Architectual Structures 1 PDF

Structures in Architecture G G Schierle
Structures in Achitecture
Excerpts
G G Schierle, PhD, FAIA
Professor of Architecture
University of Southern California
ISBN 0181965097
Copyright © G G Schierle 1990-2006. All rights reserved
This book includes material from the following sources:
International Building Code 2003:

“Portions of this document reproduce sections from the 2003 International Building Code, University of Southern California
International Code Council, Falls Church, Virginia. All rights reserved.” Custom Publishing
American Institute of Steel Construction: C/O Chancey Jemes
“Copyright © American Institute of Steel Construction, Inc. Los Angeles, CA 90089-2540
Reprinted with permission. All rights reserved.” e-mail: jemes@usc.edu
Tel. 213-740-8946
United States Geological Survey, courtesy USGS Fax 213-740-7686
About the book Acknowledgements
Structures not only support gravity and other loads, but are essential to define form and I would like to thank the following students and professionals for contributions. Students:
space. To design structures in synergy with form and space requires creativity and an Bronne Dytog, June Yip, Othman Al Shamrani, Lauramae Bryan, Sabina Cheng, Xiaojun
informed intuition of structural principles. The objective of this book is to introduce the Cheng. Claudia Chiu, Samy Chong, Kristin Donour, Miriam Figueroa, Ping Han, Lucia
principles as foundation of creative design and demonstrate successful application on Ho, Maki Kawaguchi, Nick Ketpura, Ping Kuo, Jennifer Lin, Jason Mazin, Sassu Mitra,
many case studies from around the world. Richly illustrated, the book clarifies complex Rick Patratara, Timothy Petrash, Musette Profant, Katie Rahill, Shina and Srinivas Rau,
concepts without calculus yet also provides a more profound understanding for readers Neha Sivaprasad, Madhu Thangavelu, Sharmilla Thanka, Reed Suzuki, Bogdan
with an advanced background in mathematics. The book also includes structural details Tomalevski, Carole Wong, Nasim Yalpani, Matt Warren; professionals: James Ambrose,
in wood, steel, masonry, concrete, and fabric to facilitate design of structures that are Julie Mark Cohen, Jeff Guh, Robert Harris, Theo Heizmann, Will Shepphird, Robert
effective and elegant. Many graphs streamline complex tasks like column buckling or Timme, Helge Wang, Walter Winkle; drawings by architects and engineers: Kurt
design for wind and seismic forces. The graphs also visualize critical issues and Ackerman, Ove Arup, Tigran Ayrapetian, Fred Basetti, Brabodh Banvalkar, Mario Botta,
correlate US with metric SI units of measurement. These features make the book useful Andrea Cohen Gehring, Jacques de Brer, Norman Foster, Arie Krijgsman, Von Gerkan
as reference book for professional architects and civil engineers as well as a text book Marg, David Gray, Jürgen Hennicke, Heinz Isler, Arata Isozaki, Paul Kaufmann, Pierre
for architectural and engineering education. The book has 612 pages in 24 chapters. Koenig, Panos Koulermos, Robert Marquis, Edward Niles, Frei Otto, John Portman, Jörg
Schlaich, Peter von Seidlein, James Tyler, and Dimitry Vergun.
To My Family
Units Prefixes
SI * units (metric) Conversion US units Prefix Factor
Remark factor ** Remark
Micro- 0.000001
Length
Millimeter mm 25.4 Inch in MIli-, m 0.00001
Centimeter cm 10 mm 30.48 Foot ft 12 in Centi- 0.01
Meter m 1000 mm 0.9144 Yard yd 3 ft Deci- 0.1
Kilometer km 1000 m 1.609 Mile mi 5280 ft Semi-, hemi-, demi- 0.5
Area Uni- 1
Square millimeter mm2 645.16 Square in in2 Bi-, di- 2
Sq. centimeter cm2 100 mm2 929 Square foot ft2 144 in2 Tri-, ter- 3
Square meter m2 1 Mil 0.835 Sq. yard yd2 9 ft2 Tetra-, tetr-, quadr- 4
Hectar ha 10000 m2 2.472 Acre Acre = 4840 yd2
Pent-, penta-, quintu- 5
Volume
Cubic millimeter mm3 16387 Cubic inch in3 Sex-, sexi-, hexi-, hexa-, 6
Cubic centimeter cm3 1 k mm3 28317 Cubic foot ft3 Hep-, septi-, 7
Cubic meter m3 1 Mil cm3 0.7646 Cubic yard yd3 Oct-, oct-, octa-, octo- 8
Liter l 0.001 m3 0.264 Gallon US gal = 3.785 liter Non-, nona- 9
Mass Dec-, deca-, deci, deka- 10
Gram g 28.35 Ounce oz Hect-, hector- 100
Kilogram kg 1000 g 0.4536 Pound Lb, # 16 oz Kilo-, k 1,000
Tonn t 1000 kg 0.4536 Kip k 1000 # Mega-, M 1,000,000
Force / load
Giga-, G 1,000,000,000
Newton N 4.448 Pound Lb, #
Kilo Newton kN 1000 N 4.448 Kip k 1000 #
Tera- 1,000,000,000,000
Newton/ meter N/m 14.59 Pound/ ft plf
Kilo Newton/ m kN/m 14.59 Kip/ ft klf 1000 plf
Stress
Pascal= N/m2 Pa 6895 Pound/ in2 psi
Kilo Pascal kPa 1000 Pa 6895 Kip / in2 ksi 1000
Fabric stress
Newton / m N/m 175 Pound/ in Lb/in Fabric
Load / soil pressure
Pascal Pa 1000 Pa 47.88 Pound/ ft2 psf
Moment
Newton-meter N-m 1.356 Pound-foot Lb-ft, #’
Kilo Newton-m kN-m 1000 N- 1.356 Kip-foot k-ft, k’ 1000#’
Temperature
Celcius °C .55(F-32) Fahrenheit °F
Water freezing 0°C = 32°F
Water boiling 100°C = 212°F
* SI = System International (French - designation for metric system)
** Multiplying US units with conversion factor = SI units
Dividing SI units by conversion factor = US units
5 Strength Stiffness Stability
Contents 5-2 Force types
PART I: BACKGROUND 5-3 Force vs. stress
1 Historic Evolution 5-4 Allowable stress
1-2 Walls 5-5 Axial stress
1-6 Post-and-beam 5-6 Shear stress
1-10 Arch, Vault, Dome 5-8 Torsion
1-21 Suspended 5-9 Principal stress
1-24 Truss 5-10 Strain
1-26 Skyscraper 5-10 Hook’s law
5-11 Elastic Modulus
2 Loads 5-14 Thermal strain
2-2 Introduction 5-14 Thermal stress
6
2-2 Dead load 5-17 Stability
2-4
2-5
Live load
Seismic load 6
6-4
Bending
Bending and shear
0 - 2 0 0
9
2-6 Wind load
9
6-8 Equilibrium method
1
2-8 Tributary load and load path
erle
6-10 Area method
3 Basic Concepts 6-13 Indeterminate beams
3-2 Synergy, Strength, Stiffness, Stability
c h i
6-14 Flexure formula
GS
3-3 Rupture length 6-15 Section modulus
3-4 Horizontal structures 6-16 Moment of inertia
G
Slab, plate, deck (one & two-way) 6-18 Shear stress
Beam, arch and cable
Truss
r i g h t 6-22 Deflection
3-9
Cop
Wall
Cantilever
Moment frame
y
Vertical/lateral structures 7
7-3
7-3
7-4
7-5
Buckling
Euler formula
Slenderness ratio
Combined stress
Kern
Braced frame
7-6 Arch and vault
PART II: MECHANICS 7-7 Wood buckling
4 Statics 7-12 Steel buckling
4-2 Force and moment
4-3 Static equilibrium PART III: DESIGN METHODS
4-4 Supports 8 ASD, LRFD, Masonry and Concrete Design
4-5 Reactions 8-2 ASD (Allowable Stress Design)
4-10 Static determinacy 8-3 LRFD (Load Resistance Factor Design)
4-13 Vector analysis 8-4 Masonry design (ASD)
4-15 Truss analysis 8-10 Concrete strength design (LRFD)
4-17 Funicular
4-21 Vector reactions
9 Lateral Force Design 13 Form-Resistant
9-2 Design for wind 13-2 Funicular concepts
9-8 Seismic design 13-4 Arch
9-13 SD-graphs 13-10 Vault
9-15 Analysis steps 13-17 Dome
9-18 Vertical distribution 13-23 Grid shell
9-19 Horizontal diaphragms 13-29 HP shell
9-22 Eccentricity 13-37 Freeform shell
9-23 Hazard configurations
9-24 Stability issues 14 Tensile Resistant
9-27 Seismic safety items 14-1 Tension members
14-2 Prestress
10 Conceptual Design 14-3 Stayed structures
6
10-1 System selection 14-8 Propped structures
10-3
10-4
10-5
Global moment and shear
Radial pressure
Examples
14-10
14-17
14-21
Suspended structures
Cable truss
Anticlastic structures
0 - 2 0 0
10-7 Case studies 14-42
9
Pneumatic structures
1 9
erle
10-15 Portal method
10-17 Moment frame PART V: VERTICAL STRUCTURES
10-19 Braced frame
c h i
GS
10-21 Test models 15 General Background
10-23 Sample projects 15-2 Tall structures
G
10-29 Computer aided design 15-3 Gravity load
PART IV: HORIZONTAL SYSTEMS
r i g h t 15-4
15-7
Lateral load
Structure systems
11
11-1
11-3
11-5 Cop
Bending Resistant
Bending concepts
Beam optimization
y
Joist, beam, girder
15-11
15-12
16
16-2
Floor framing
Beam-column interaction
Shear Resistant
Classic walls
11-11 Vierendeel 16-3 Seismic failures
11-17 Folded plate 16-4 Shear walls
11-22 Cylindrical shell 16-6 Shear wall stability
16-7 Wood shear walls
12 Axial Resistant 16-10 Shear wall reinforcing
12-2 Truss
Truss configurations 17 Bending Resistant
Prismatic truss 17-2 Cantilever
Folded truss 17-6 Moment frame
12-13 Space truss 17-13 Framed tube
12-22 Tree structures 17-16 Bundled tube
18 Axial Resistant 24 Cable and Fabric
18-2 Braced frame 24-1 Material
18-8 Belt truss and outrigger 24-2 Fabric
18-12 Braced tube 24-4 Cables
18-16 Eccentric braced frame 24-10 Projects
19 Suspended high-rise Appendix A: Beam Formulas

19-2 Suspension rational A-2 Beam formulas
19-3 Design options A-3 Bending coefficients
19-3 Limits
19-4 Case studies Appendix B: Geometric Properties
B-2 Centroid
PART VI: MATERIAL B-4 Moment of Inertia
6
B-6 Parallel Axis Theorem
20
20-1
20-5
Wood
Material
Heavy timber
B-7
B-8
Radius of Gyration
Geometric properties
0 - 2 0 0
20-13 Grid structures
1 9
Appendix C: Lateral Design Data
9
erle
Balloon framing C-2 Wind design data
Platform framing C-7 Seismic design data
20-29 Projects
c h i
GS
Appendix D: Material and Buckling Data
21 Steel D-2 Wood
G
21-1 Material D-8 Steel
21-7
21-29
Heavy steel
Light gauge steel
rig h t Appendix E: Design Tables

21-33
22
22-1
22-7 C o p y
Projects
Masonry
Material
Brick masonry
E-2
E-3
Index
Span Ranges for Structure Elements
Span Ranges for Structure Systems
22-18 Concrete masonry

22-22 Stone masonry
22-23 Projects
23 Concrete
23-1 Material
23-4 Reinforced concrete
23-17 Prestressed concrete
23-20 Precast concrete
23-24 Tilt-up concrete
23-26 Projects
Understanding loads on buildings is essential for structural design and a major factor to
2 define structural requirements. Load may be static, like furniture, dynamic like
earthquakes, or impact load like a car hitting a building. Load may also be man-made,
like equipment, or natural like snow or wind load. Although actual load is unpredictable,
design loads are usually based on statistical probability. Tributary load is the load
imposed on a structural element, like a beam or column, used to design the element. All
of these aspects are described in this chapter.
Loa d
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop
2-1 BACKGROUND Load

Introduction
Structures resist various loads (gravity, seismic, wind, etc.) that may change over time.
For example, furniture may be moved and wind may change rapidly and repeatedly..
Loads are defined as dead load (DL) and live load (LL); point load and distributed
load; static, impact, and dynamic load, as shown at left.
1 Dead load: structure and permanently attached items (table 21.)
2 Live load: unattached items, like people, furniture, snow, etc (table 2.2)
3 Distributed load (random – snow drift, etc.)
4 Uniform load (uniform distribution)
5 Point load (concentrated load)
6 Uniform load on part of a beam is more critical than full load
7 Negative bending over support under full load reduces positive bending
8 Static load (load at rest)
9 Impact load (moving object hitting a structure)
2 0 0 6
10 Dynamic load (cyclic loads, like earthquakes, wind gusts, etc.)
-
•
1 9 9 0
Classification as DL and LL is due to the following considerations:
Seismic load is primarily defined by dead load
erle
• Dead load can be used to resist overturning under lateral load
•
i
Long term DL can cause material fatigue
GS c h•
•
DL deflection may be compensated by a camper (reversed deflection)
For some elements, such as beams that span more than two supports
partial load may be more critical than full load; thus DL is assumed on the
h t G full beam but LL only on part of it
rig
Lateral load (load that acts horizontally) includes:
C o p y • Seismic load (earthquake load)

• Wind load
• Soil pressure on retaining walls
Other load issues introduced::

• Tributary load (load acting on a given member)
• Load path (the path load travels from origin to foundation)
Dead Load
Dead load is the weight of the structure itself and any item permanently attached to it,
Dead load defines the mass of buildings for seismic design. Table 2.1 give the weight of
materials to define building mass. Approximate dead loads are:
• Wood platform framing: 14 psf
• Wood platform framing with lightweight concrete: 28 psf
• Steel framing with concrete deck: 94 to 124 psf
2-2 BACKGROUND Load

Table 2.1. Material weight Table 2.1. Material weight - continued
Weight by volume US units SI units Weight by area psf Pa
Masonry / concrete / etc. pcf kg/dm3 Gypsum board, 5/8” (16 mm) 2.5 120
Brick 120 1.92 Stucco, 7/8” (22 mm) 8 383
Concrete masonry units (CMU) 100 1.60 Acoustic tile, ½” 0.8 38
Light-weight CMU 60 0.96 Ceramic tile, 1/4” (6.3 mm) 2.5 120
Concrete 150 2.40 Glass
Vermiculite concrete 25 - 60 0.40-0.96 Sheet glass, 1/8” (3 mm) 1.5 72
Gravel / sand 90-120 1.44-1.92 Sheet glass, ¼” (6 mm) 3 144
Soil
Water at 4° C
75-115
62.4
1.20-1.84
1.00
Glass block, 4” (102 mm)
Roof material
- 2 0 0
20
6 958
Metals
Aluminum 165 2.64
Built-up roof
1 9
Clay / concrete tiles
9 0 7
13-20
335
622-959
erle
Cast iron 450 7.21 Metal 1-3 48-144
Steel 485
c h i7.77 Shingles 3 144
GS
Stainless steel 492-510 7.88-8.17 Single ply roof / fabric roof 1-2 48-06
Copper
h t G 556 8.91 Industrial fabric (PVC , fiber glass) 1-2 48-96
rig
Lead 710 11.38 Steel floor constructions
Stone
C o p y
Granite / slate
Lime stone / marble
175
165
2.81
2.64
Steel deck / concrete slab, 6” (15 mm)
Suspended ceiling
Floor finish
40-60
2
2
1915-2873
96
96
Sandstone 150 2.40 Steel framing (varies with height) 10-40 479-1915
Wood Partitions (required by code) 20 958
Cedar 22 0.35 Total 94 - 124 3543-5458
Douglas fir 34 0.55 Wood platform framing
Oak 47 0.75 Wood platform framing + floor / ceiling 14 670
Pine, white 25 0.40 Light-weight concrete option 14 670
Redwood 28 0.45 Total (with and without concrete) 14 - 28 1341
2-3 BACKGROUND Load

Live Load
IBC table 1607.1 excerpts. Minimum uniform live load IBC table 1607.1 defines live loads for various occupancies. Except for live load >100 psf
Use or Occupancy Pounds / ft2 kPa (4.79 kPa) these loads may be reduced for large tributary areas as follows:
Category Description psf kPa
R = r (A - 150)
Office use 50 2.39 R = r (A – 14) [for SI units]
Access floors
Computer use 100 4.79 Reductions R shall not exceed
Fixed seating 50 2.39 • 40% for horizontal members
Auditoria and • 60 % for vertical members
Movable seating 100 4.79 • R = 23.1 (1+ D/L)
Assembly areas
Stage 125 5.99 where
R = reduction in percent
Storage and repair 100 4.79
Garages r = 0.08 for floors
Hospitals
Private
Wards and rooms
50
40
2.39
1.92
A = tributary area in square foot (m2)
D = dead load
- 2 0 0 6
0
L = Unreduced live load per square foot (m2)
9
Reading room 60 2.87
Libraries
Stack room 125 5.99 Roof Load
1 9
erle
Light 75 3.59 Roof loads are defined by IBC
Manufacturing
Heavy 125
c h i
5.99
• Wind load per IBC 1609
GS
• Snow load per IBC 1608
Offices 50 2.39 • Minimum roof loads:
Printing plants
h t G
Press room 150 7.18 Roof type
Awnings and canopies
psf
5
Pa
240
rig
Composing, etc. 100 4.79
Green houses 10 479
o p
Residential
C y
Reviewing stands, etc..
Basic floor area
Exterior balconies
40
60
100
1.92
2.87
4.79
Landscaped roofs (soil + landscaping as DL)
General flat, pitched, and curved roofs
Lr = 20R1 / R2
where
20
Lr
958
Lr
Schools Classrooms 40 1.92 12 < Lr < 20

Sidewalks and driveways Public access 250 11.97 0.58 < Lr < 0.96 for SI units
R1 = 1 for A ≤ 200 sq. ft. for A ≤ 19 m2
Light 125 5.99 R1 = 0.6 for A ≥ 600 sq. ft. for A ≥ 56 m2
Storage
Heavy 250 11.97 R2 = 1 for slopes ≤ 1:3 ≤ 18°
R2 = 0.6 for slopes ≥ 1:1 ≥ 45°
Stores 100 4.79
A = tributary area
Pedestrian bridges 100 4.79 See IBC1607.11 for other R1 and R2 factors
2-4 BACKGROUND Load

Seismic load
Earthquakes cause horizontal and vertical ground shaking. The horizontal (lateral)
shaking is usually most critical on buildings. Earthquakes are caused by slippage of
seismic fault lines or volcanic eruption. Fault slippage occurs when the stress caused by
differential movement exceeds the soil shear capacity. Differential movement occurs
primarily at the intersection of tectonic plates, such as the San Andreas fault which
separates the pacific plate from the US continental plate. Earthquake intensity is
greatest after a long accumulation of fault stress. Seismic waves propagate generally in
radial patterns, much like a stone thrown in water causes radial waves. The radial
patterns imply shaking primarily vertical above the source and primarily horizontal with
distance. The horizontal shaking usually dominates and is most critical on buildings.
Although earthquakes are dynamic phenomena, their effect mat be treated as equivalent
static force acting at the base of buildings. This lateral force, called base shear, is
basically governed by Newton’s law:
f=ma (force = mass x acceleration)
- 2 0 0 6
9 9 0
Base shear is dampened by ductility, a structure’s capacity to absorb energy through
1
elastic deformation. Ductile structures deform much like flowers in the wind, yet brittle
erle
(non-ductile) structures sustain greater inertia forces. Steel moment resisting frames are
i
ductile, though some shear walls are brittle. In earthquake prone areas seismic base
GS c hshear as percentage of mass is approximately:

• ~ 4 % for tall ductile moment frames
i g h t G •
•
~ 10 % for low-rise ductile moment frames
~ 15 % for plywood shear walls
y r • ~ 20 to 30 % for stiff shear walls
Cop
Seismic design objectives:
• Minimize mass
• Maximize ductility
1. Fault rupture / wave propagation
(predominant vertical above rupture, lateral at distance)
2 Lateral slip fault
3 Thrust fault
4 Building overturning
5 Base shear
6 Bending deformation (first mode)
7 Bending deformation (higher mode)
E Epicenter (earthquake source above ground)
H Hyper center (actual earthquake source under ground)
2-5 BACKGROUND Load

Wind load (see also Lateral Force Design)
Wind load generates lateral forces, much like earthquakes. But, though seismic forces
are dynamic, wind load is usually static, except gusty wind and wind on flexible
structures. In addition to pressure on the side facing the wind (called wind side), wind
also generates suction on the opposite side (called lee side) as well as uplifting on roofs.
Wind pressure on buildings increases with increasing velocity, height and exposure. IBC
Figure 16-1 gives wind velocity (speed). Velocity wind pressure (pressure at 33 feet, 10
m above the ground) is defined by the formula
q = 0.00256 V2 (H/33)2/7
Where
q = velocity pressure in psf [1 psf = 47.9 Pa]
V = velocity in mph [1 mph = 1.609 km/h]
H = height in feet
- 2 0 0 6
[1 ft = 0.305 m]
The actual wind pressure P is the velocity pressure q multiplied by adjustment factors
1 9 9 0
based on empirical data from wind tunnel tests, tabulated in code tables. The factors
account for type of exposure, orientation, and peak pressure along edges, roof ridge, and
erle
method of analysis. IBC defines three exposures and two methods:
c h
•
•
i Exposure B
Exposure C
(sites protected by buildings or a forest)
(open sites outside cities)
GS
• Exposure D (sites near an ocean or large lake)
G
• Method 1: Normal force method
r i g h t • Method 2: Projected area method
y
Depending on location, height, and exposure, method 2 pressures range from 10-110 psf
Cop
(0.5 to 5 kPa). This is further described in Lateral Force Design
Design objectives for wind load:
• Maximize mass to resist uplift
• Maximize stiffness to reduce drift
1 Wind load on gabled building (left pressure, right suction)
2 Wind load on dome or vault (left pressure, right suction)
3 Buildings within cities are protected by other buildings
4 Tall building exposed to full wind pressure
5 Wind on wide façade is more critical than on narrow facade
6 Building forms increase wind speed
2-6 BACKGROUND Load

Tributary load and load path
Tributary load is the load acting on any element, like a beam, column, slab, wall,
foundation, etc. Tributary load is needed to design / analyze any element.
Load path is the path any load travels from where it originates on a structure to where it
is ultimately resisted (usually the foundation). It is essential to define the tributary load.
The following examples illustrate tributary load and load path
6
1 Simple beam / 2 columns
Assume
Uniform beam load w = 200 plf
0 - 2 0 0
9
Beam span L = 30’
Find
1 9
erle
Load path: beam / column
Tributary load: Reactions at columns A and B
c h i Ra = Rb = R = w L/2
GS
R = 200x30/2 = 3000#
Convert pounds to kip
G
R = 3000#/1000 R = 3.0 k
r i g h t 2 Two beams / three columns

Assume
Copy Uniform beam load w = 2 klf

Beam spans L1 = 10’, L2 = 20’
Find
Load path: beam / column
Tributary load: Reactions at columns A, B, C
Ra = 2x10/2 Ra = 10 k
Rb = 2(10+20)/2 Rb = 30 k
Rc = 2x20/2 Rc = 20 k
2-7 BACKGROUND Load

1 One-story concrete structure
Assume
Roofing 3 psf
Ceiling 2 psf
10” con. slab 125 psf (150 pcf x10”/12”)
DL 130 psf
LL 20 psf
Σ 150 psf
Lx = 30’, Lxc = 34’, Ly = 25’
Columns, 12”x12” (t=12”, t/2 = 6” = 0.5’)
Column reactions A, B, C, D
Ra = 150 psf (30+34)/2 (25) Ra = 120,000 #
Rb = 150 (30+34)/2 (25/2+0.5) Rb = 62,400 #
6
Rc = 150 (30/2+0.5) (25) Rc = 58,125 #
2
Rd = 150 (30/2+0.5) (25/2+0.5)
Three-story concrete structure
0 - 2 0 0 Rd = 30,225 #
Assume
1
Roof DL
9 130 psf
9
erle
Roof LL 20 psf
i
Roof Σ 150 psf
GS c h Floor DL
Floor LL
150 psf
50 psf
(includes columns, etc.)
(Office)
i g h t G Floor Σ 200 psf

Lx = 30’, Lxc = 34’, Ly = 25’
r
Columns, 2’x2’ (t =2’, t/2 =1’)
Copy Column reactions at level 2

Ra = 150 psf (30+34)/2 (25) =
Rb = 150 (30+34)/2 (25/2+1) =
Rc = 150 (30/2+1) (25) =
150 x 800
150 x 432
150 x 400
w = 150 psf)
Ra = 120,000 #
Rb = 64,800 #
Rc = 60,000 #
Rd = 150 (30/2+1) (25/2+1) = 150 x 216 Rd = 32,400 #
Column reactions at level 1, w = 150+200 w = 350 psf
Ra = 350 (800) Ra = 280,000 #
Rb = 350 (432) Rb = 151,200 #
Rc = 350 (400) Rc = 140,000 #
Rd = 350 (216) Rd = 75,600 #
Column reactions at level 0, w = 150+200+200 w = 550 psf
Ra = 550 (800) Ra = 440,000 #
Rb = 550 (432) Rb = 237,600 #
Rc = 550 (400) Rc = 220,000 #
Rd = 550 (216) Rd = 118,800 #
2-8 BACKGROUND Load

Deck / joist / beam / column
Assume
Uniform load w = 80 psf
Joist spacing e = 2’
Joist span L1 = 12’
Beam spans
L2 = 10’
L3 = 20’
Find load path and tributary load
Load path: plywood deck / joist / beam / columns
6
Tributary loads:
Uniform joist load
wj = w e = 80 psf x 2’
0-2 0 0
Beam load (assume uniform load due to narrow joist spacing)
wj = 160 plf
9 9
wb = 80 psf L1/2 = 80 psf x 12’ /2
1
wb = 480 plf
erle
Column reaction
Ra = wb L2 / 2 = 480 plf x 10 /2 Ra = 2,400 #
c h i
Rb = wb (L2+L3)/2 = 480 (10+20) / 2 Rb = 7,200 #
GS
Rc = wb L3 / 2 = 480 x 20 / 2 Rc = 4,800 #
i g h t G
y r
Cop
2-9 BACKGROUND Load

Concrete slab / wall / footing / soil
Allowable soil pressure 2000 psf (for stiff soil)
Concrete slab, 8” thick
Slab length L = 20’
Wall height h = 10’
DL = 100 psf (150 pcf x 8”/12”)
6
LL = 40 psf (apartment LL)
Σ = 140 psf
8” CMU wall, 10’ high at 80 psf
0 - 2 0 0
9
(CMU = Concrete Masonry Units)
1
(8” nominal = 7 5/8” = 7.625” actual)
9
erle
Concrete footing 2’ x 1’at 150 pcf
i
nalyze a 1 ft wide strip (1 meter in SI units)
GS c hSlab load
w = 140psf x 20 / 2 w = 1400 #
G
CMU wall load
rig h t w = 80 psf x 10’

Footing load
w = 800 #
Co p y w = 150 pcf x 2’x 1’

Total load on soil
P = 1400 + 800 + 300
Soil pressure
w = 300 #
P = 2500 #
f = P/A = 2500 # / (1’x2’) f =1250 psf

1250 < 2000, ok
2-10 BACKGROUND Load

1 Concrete slab / wall
Concrete slab t = 8”, span L = 20’
LL = 50 psf
DL = 120 psf
Σ = 170 psf
Slab load on wall
w = 170 psf x 20’/2 w =1700 plf
2 Joist roof / wall
Plywood roof deck, 2x12 wood joists at 24”, span L = 18’
LL = 30 psf
DL = 20 psf
Σ = 50 psf
Roof load on wall (per linear foot of wall length)
6
w = 50 psf x 18’/2 w = 450 plf
3 Concrete slab / beam / wall
0 - 2 0 0
Concrete slab t = 5”, span L =10’, beam span L = 30’
LL =
DL =
1 9
20 psf
9
70 psf (assume beam DL lumped with slab DL)
erle
Σ = 90 psf
i
Beam load w = 90 psf x10’ / 1000 w = 0.9 klf
GS c h4
Wall reaction R = 0.9 klf x 30’ / 2
Concrete slab on metal deck / joist/ beam
Spans: deck L = 8’, joist L = 20’, beam L=40’
R = 13.5 k
i g h t G LL =
DL =
40 psf
60 psf (assume joist and beam DL lumped with slab DL)
y r Σ = 100 psf
Cop
Joist load w = 100 psf x 8’ / 1000 w = 0.8 klf
Beam point loads P = 0.8 klf x 20’ P = 16 k
Wall reaction R = 4 P /2 = 4 x 16 k / 2 R = 32 k
Note: wall requires pilaster to support beams
5 Concrete slab on metal deck / joist/ beam / girder
Spans: deck L = 5’, joist L = 20’, beam L=40’, girder L = 60’
LL = 50 psf
DL = 50 psf (assume joist/beam/girder DL lumped with slab DL)
Σ =100 psf
Uniform joist load w = 100 psf x 5’/1000 w = 0.5 klf
Beam point loads P = 0.5 klf x 20’ P = 10 k
Girder point loads P = 7 x 10 k/2 P = 35 k
Column reaction R = (100 psf/1000) x 40’ x 60’/4 R = 60 k

Wind load resisted by shear wall
1 Three-story building
2 Exploded visualization
3 Dimensions
A Wind Wall
B Diaphragms
C Shear walls
Assume:
Building dimensions as shown in diagram
Wind pressure P = 20 psf
Find load path and tributary load
6
Load path
Wind wall > diaphragms > shear walls > footings
0 - 2 0 0
9
Note:
9
Floor and roof diaphragms act like beams to transfer load from wind wall to shear wall
1
erle
Tributary loads
c h i
Roof diaphragm
GS
V3 = 20 psf x 100’ x 5’/1000 V3 = 10 k
Level 2 diaphragm
G
V2 = 20 psf x 100’ x 10’/1000 V2 = 20 k
rig h t Level 1 diaphragm

V1 = 20 psf x 100’ x 10’/1000 V1 = 20 k
C o p y Shear walls
Level 2 shear walls
V2 = 10 k / 2
Level 1 shear wall
V2 = 5 k
V = (10 k + 20) k / 2 V1 = 15 k
Level 0 shear walls
V0 = (10 k + 20 k + 20) k / 2 V0 = 25 k
Note:
• Floor and roof diaphragms resist half the load from above and below
• Floor and roof diaphragms transfer load from wind wall to shear walls
• The 2 shear walls resist each half of the diaphragm load from above

This chapter on basic concept introduces:
3 • Structural design for:

• Strength
• Stiffness
• Stability
• Synergy
• Rupture length (material properties, i.e., structural efficiency)

Ba sic Conc e pt s
• Basic structure systems
• Horizontal structures
• Vertical / lateral structures for:
o Gravity load
- 2 0 0 6
0
o Lateral load
1 9 9
c h i erle
h t G GS
y r i g
Cop
3-1 BACKGROUND Basic Concepts

Strength, Stiffness, Stability, Synergy
Structures must be designed to satisfy three Ss and should satisfy all four Ss of structural
design – as demonstrated on the following examples, illustrated at left.
1 Strength to prevent breaking
2 Stiffness to prevent excessive deformation
3 Stability to prevent collapse
4 Synergy to reinforce architectural design, described on two examples:
Pragmatic example: Beam composed of wooden boards
Philosophical example: Auditorium design
defined by the Moment of Inertia, I = b d3 / 12
- 2 0 0 6
Comparing beams of wooden boards, b = 12” wide and d = 1”deep, each. Stiffness is
1 board, I = 12x13/12
1 9
10 boards I = 10 (12x13/12)
9 0 I=1
I = 10
erle
10 boards glued, I = 12x103/12 I = 1000
c h i
Strength is defined by the Section modulus, S = I / (d/2)
GS
1 board, S = 1/o.5 S=2
G
10 boards, S = 10/0.5 S = 20
r i g h t 10 boards, glued, S =1000/5 S = 200
Copy Note:
The same amount of material is 100 times stiffer and 10 times stronger when glued
together to transfer shear and thereby engage top and bottom fibers in compression and
tension (a system, greater than the sum of its parts). On a philosophical level, structures
can strengthen architectural design as shown on the example of an auditorium:
• Architecturally, columns define the circulation
• Structurally, column location reduces bending in roof beams over 500% !

Rupture length
Rupture length is the maximum length a bar of constant cross section area can be
suspended without rupture under its weight in tension (compression for concrete &
masonry).
Rapture length defines material efficiency as strength / weight ratio:
R=F/λ
R = rupture length
F = breaking strength
λ = specific gravity (self weight)
0 0 6
Rupture length, is of particular importance for long-span structures. The depth of
- 2
horizontal span members increases with span. Consequently the weight also increases
1 9 9 0
with span. Therefore the capacity of material to span depends on both its strength and
weight. This is why lightweight material, such as glass fiber fabrics are good for long-
erle
span structures. For some material, a thin line extends the rupture length to account for
i
different material grades.
GS c h
i g h t G
y r
Cop
The graph data is partly based on a study of the Light weight Structures Institute,
University Stuttgart, Germany

Horizontal structures
Horizontal systems come in two types: one way and two way. Two way systems are only
efficient for spaces with about equal span in both directions; as described below. The
diagrams here show one way systems at left and two way systems at right
1 Plywood deck on wood joists
2 Concrete slab on metal deck and steel joists
3 One way concrete slab
4 One way beams
5 One way rib slab
6 Two way concrete plate
7 Two way concrete slab on drop panels
8 Two way concrete slab on edge beams
6
9 Two way beams
10
11
Two way waffle slab
Deflection Δ for span length L1
0 - 2 0 0
9
12 Deflection Δ=16 due to double span L2 = 2 L1
Note:
1 9
erle
Deflection increases with the fourth power of span. Hence for double span deflection
c h i
increase 16 times. Therefore two way systems over rectangular plan are ineffective
because elements that span the short way control deflection and consequently have to
GS
resist most load and elements that span the long way are very ineffective.
i g h t G
y r
Cop
11 12

Trusses
Trusses support load much like beams, but for longer spans. As the depth and thus
dead weight of beams increases with span they become increasingly inefficient, requiring
most capacity to support their own weight rather than imposed live load. Trusses replace
bulk by triangulation to reduce dead weight.
1 Unstable square panel deforms under load.
Only triangles are intrinsically stable polygons
2 Truss of triangular panels with inward sloping diagonal bars
that elongate in tension under load (preferred configuration)
3 Outward sloping diagonal bars compress (disadvantage)
4 Top chords shorten in compression
2 0 6
Bottom chords elongate in tension under gravity load
0
Gable truss with top compression and bottom tension
-
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Warren trusses
Pompidou Center, Paris by Piano and Rogers
- 2 0 0 6
1 9 9 0
erle
Prismatic trusses
c h i
IBM Sport Center by Michael Hopkins
GS
(Prismatic trusses of triangular cross section provide rotational resistance)
i g h t G
y r
Cop
Space trusses
Square and triangular plan
Note:
Two way space trusses are most effective if the spans in the principle directions are
about equal, as described for two-way slabs above. The base modules of trusses should
be compatible with plan configuration (square, triangular, etc.)

Funicular structures
The funicular concept can be best described and visualized with cables or chains,
suspended from two points, that adjust their form for any load in tension. But funicular
structures may also be compressed like arches. Yet, although funicular tension structures
adjust their form for pure tension under any load, funicular compression structures may
be subject to bending in addition to compression since their form is rigid and not
adaptable. The funicular line for tension and compression are inversely identical; the
form of a cable becomes the form of an arch upside-down. Thus funicular forms may be
found on tensile elements.
1 Funicular tension triangle under single load
2 Funicular compression triangle under single load
3 Funicular tension trapezoid under twin loads
4
5
2
Funicular tension polygon under point loads
- 0 0
Funicular compression trapezoid under twin loads
6
0
6 Funicular compression polygon under point load
7
8
9 9
Funicular tension parabola under uniform load
1
Funicular compression parabola under uniform load
c h i erle
h t G GS
y r i g
Cop

Vault
IBM traveling exhibit by Renzo Piano
A series of trussed arches in linear extrusion form a vault space. The trussed arches
consist of wood bars with metal connectors for quick assembly and disassembly as
required for the traveling exhibit. Plastic panels form the enclosing skin. The trussed
arches provide depth and rigidity to accommodate various load conditions
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop Suspension roof
Exhibit hall Hanover by Thomas Herzog

Vertical structures
Vertical elements
Vertical elements transfer load from roof to foundation, carrying gravity and/or lateral
load. Although elements may resist only gravity or only lateral load, most are designed to
resist both. Shear walls designed for both gravity and lateral load may use gravity dead
load to resist overturning which is most important for short walls. Four basic elements
are used individually or in combination to resist gravity and lateral loads
1 Wall under gravity load
2 Wall under lateral load (shear wall)
3 Cantilever under gravity load
4 Cantilever under lateral load
5
6
Moment frame under gravity load
Moment frame under lateral load
- 2 0 0 6
0
7 Braced frame under gravity load
9
9
Braced frame under lateral load
1 9
c h i erle
h t G GS
y r i g
Cop

Vertical systems
Vertical systems transfer the load of horizontal systems from roof to foundation, carrying
gravity and/or lateral load. Although they may resist gravity or lateral load only, most
resist both, gravity load in compression, lateral load in shear. Walls are usually designed
to define spaces and provide support, an appropriate solution for apartment and hotel
buildings. The four systems are:
1 Shear walls (apartments / hotels)
2 Cantilever (Johnson Wax tower by F L Wright)
3 Moment frame
4 Braced frame
A Concrete moment resistant joint
Column re-bars penetrate beam and beam re-bars penetrate column)
6
B Steel moment resistant joint
0 - 2 0 0
(stiffener plates between column flanges resist beam flange stress)
1 9 9
Vertical / lateral element selection criteria
erle
Element Advantages Challenges
Shear wall Good for Inflexible for future changes
c h i Architectural criteria apartments/hotels
GS
Stiffness increases seismic
Structural criteria Very stiff, good for forces
i g h t G Cantilever
wind resistance
Flexible planning Must remain in future
r
Architectural criteria Around cantilever changes
Copy Structural criteria
Moment frame
Ductile, much like a tree
trunk
Most flexible, good for
Too flexible for tall
structures
Expensive, drift may cause
Architectural criteria office buildings problems
Structural criteria Ductile, absorbs seismic Tall structures need

force additional stiffening
Braced frame More flexible then Less flexible than moment
B Architectural criteria Shear walls frame
Structural criteria Very stiff, good for Stiffness increases seismic

Wind resistance forces

Shear walls
As the name implies, shear walls resist lateral load in shear. Shear walls may be of
wood, concrete or masonry. In the US the most common material for low-rise
apartments is light-weight wood framing with plywood or particle board sheathing.
Framing studs, spaced 16 or 24 inches, support gravity load and sheathing resists lateral
shear. In seismic areas concrete and masonry shear walls must be reinforced with steel
bars to resist lateral shear.
1 Wood shear wall with plywood sheathing
2 Light gauge steel shear wall with plywood sheathing
3 Concrete shear wall with steel reinforcing
4 CMU shear wall with steel reinforcing
5 Un-reinforced brick masonry (not allowed in seismic areas)
8
0 6
Two-wythe brick shear wall with steel reinforcing
- 2 0
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Cantilevers
Cantilevers resist lateral load primarily in bending. They may consist of single towers or
multiple towers. Single towers act much like trees and require large footings like tree
roots to resist overturning. Bending in cantilevers increases from top down, justifying
tapered form in response.
1 Single tower cantilever
2 Single tower cantilever under lateral load
3 Twin tower cantilever
4 Twin tower cantilever under lateral load
5 Suspended tower with single cantilever
6 Suspended tower under lateral load
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Moment frames
Moment frames resist gravity and lateral load in bending and compression. They are
derived from post-and beam portals with moment resisting beam to column connections
(for convenience referred to as moment frames and moment joints). The effect of
moment joints is that load applied to the beam will rotate its ends and in turn rotate the
attached columns. Equally, load applied to columns will rotate their ends and in turn
rotate the beam. This mutual interaction makes moment frames effective to resist lateral
load with ductility. Ductility is the capacity to deform without breaking, a good property to
resist earthquakes, resulting in smaller seismic forces than in shear walls and braced
frames. However, in areas with prevailing wind load, the greater stiffness of shear walls
and braced frames is an advantage. The effect of moment joints to resist loads is
visualized through amplified deformation as follows:
1
2
2 0 0 6
Portal with pin joints collapses under lateral load
Portal with moment joints at base under lateral load
-
0
3 Portal with moment beam/column joints under gravity load
4
5
9 9
Portal with moment beam/column joints under lateral load
1
Portal with all moment joints under gravity load
erle
6 Portal with all moment joints under lateral load
c h
7
8
i High-rise moment frame under gravity load
Moment frame building under lateral load
h t G GS I
Note:
Inflection points (zero bending between negative and positive bending
p y rig Deformations reverse under reversed load
Co

Braced frames
Braced frames resist gravity load in bending and axial compression, and lateral load in
axial compression and tension by triangulation, much like trusses. The triangulation
results in greater stiffness, an advantage to resist wind load, but increases seismic
forces, a disadvantage to resist earthquakes. Triangulation may take several
configurations, single diagonals, A-bracing, V-bracing, X-bracing, etc., considering both
architectural and structural criteria. For example, location of doors may be effected by
bracing and impossible with X-bracing. Structurally, a single diagonal brace is the
longest, which increases buckling tendency under compression. Also the number of
costly joints varies: two for single diagonals, three for A- and V-braces, and five joints for
X-braces. The effect of bracing to resist load is visualized through amplified deformation
as follows:
1
2
2 0
A-braced portal under gravity and lateral load
- 0 6
Single diagonal portal under gravity and lateral loads
0
3 V-braced portal under gravity and lateral load
4
5
9 9
X-braced portal under gravity and lateral load
1
Braced frame building without and with lateral load
c h i
Note:
erle
Deformations and forces reverse under reversed load
h t G GS
y r i g
Cop

Pa rt I I 4
M e c ha nic s St a t ic s
Mechanics, as defined for the study of structures, is the behavior of physical systems Statics is the branch of mechanics that deals with forces and force systems that act on
under the action of forces; this includes both statics and dynamics.
2 0 0 6
bodies in equilibrium. Since buildings are typically designed to be at rest (in equilibrium),
the subject of this book is primarily focused on statics. Even though loads like
-
0
Dynamics is the branch of mechanics that deals with the motion of a system of material earthquakes are dynamic they are usually treated as equivalent static forces.
particles under the influence of forces. Dynamic equilibrium, also known as kinetic
equilibrium, is the condition of a mechanical system when the kinetic reaction of all forces
1 9 9
erle
acting on it is in dynamic equilibrium.
c h i
Statics is the branch of mechanics that deals with forces and force systems that act on
GS
bodies in equilibrium as described in the following.
i g h t G
y r
Cop
4-1 MECHANICS Statics

Force and Moment
Force is an action on a body that tends to:
• change the shape of an object or
• move an object or
• change the motion of an object
US units: # (pound), k (kip)
SI units: N (Newton), kN (kilo Newton)
Moment is a force acting about a point at a distance called lever arm
M =P L (Force x lever arm)
The lever arm is measured normal (perpendicular) to the force.
Moments tend to:
• rotate an object or
• bend an object (bending moment)
US units:
SI units:
- 2 0 0 6
#’ (pound-feet), k’ (kip-feet), #” (pound-inch), k” (kip-inch)
N-m (Newton-meter), kN-m (kilo-Newton-meter)
9
Gravity force (compresses the pyramid)
1 9 0
erle
2 Pulling force (moves the boulder)
3 Moment = force times lever arm (M = P L)
c hA
i Point about which the force rotates
GS
L Lever arm
M Moment
i g h t G P Force
y r
Cop

Static Equilibrium
For any body to be in static equilibrium, all forces and moments acting on it must be in
equilibrium, i.e. their sum must equal zero. This powerful concept is used for static
analysis and defined by the following three equations of statics:
ΣH=0 (all horizontal forces must equal zero)
ΣV=0 (all vertical forces must equal zero)
ΣM=0 (all moments must equal zero)
The equilibrium equations are illustrated as follows:
1 Horizontal equilibrium: pulling left and right with equal forces,
mathematically defined as
Σ H = 0 = + 100 – 100 = 0
2
mathematically defined as:
Σ V = 0 = – 2 x 100 + 200 = 0
- 2 0 6
Vertical equilibrium: pushing up with a force equal to a weight,
0
3
9 0
Moment equilibrium: balancing both sides of a balance board,
1 9
mathematically defined as:
erle
Σ M = 0 = – 50# (8’) + 200# (2’) = - 400 + 400 = 0
c h i
GS
Much of this book is based on the three equilibrium equations.
i g h t G
y r
Cop

Supports
For convenience, support types are described for beams, but apply to other horizontal
elements, like trusses, as well. The type of support affects analysis and design, as well
as performance. Given the three equations of statics defined above, ΣH=0, ΣV=0, and
ΣM=0, beams with three unknown reactions are considered determinate (as described
below) and can be analyzed by the three static equations. Beams with more than three
unknown reactions are considered indeterminate and cannot be analyzed by the three
static equations alone. A beam with two pin supports (1 has four unknown reactions, one
horizontal and one vertical reaction at each support. Under load, in addition to bending,
this beam would deform like a suspended cable in tension, making the analysis more
complex and not possible with static equations.
By contrast, a beam with one pin and one roller support (2) has only three unknown
6
reactions, one horizontal and two vertical. In bridge structures such supports are quite
- 2 0 0
common. To simplify analysis, in building structures this type of support may be
assumed, since supporting walls or columns usually are flexible enough to simulate the
0
9
same behavior as one pin and one roller support. The diagrams at left show for each
1 9
support on top the physical conditions and below the symbolic abstraction.
erle
1 Beam with fixed supports at both ends subject to bending and tension
c h
2
3
i Simple beam with one pin and one roller support subject to bending only
Beam with flexible supports, behaves like a simple beam
h t G GS Simple beams, supported by one pin and one roller, are very common and easy to
analyze. Designations of roller- and pin supports are used to describe the structural
g
behavior assumed for analysis, but do not always reflect the actual physical support. For
y r i example, a pin support is not an actual pin but a support that resists horizontal and
Cop
vertical movement but allows rotation. Roller supports may consist of Teflon or similar
material of low friction that allows horizontal movement like a roller.

Support symbols
The diagrams show common types of support at left and related symbols at right. In
addition to the pin and roller support described above, they also include fixed-end
support (as used in steel and concrete moment frames, for example).
Support types
Degrees of freedom
Support type Horizontal Vertical Rotation
movement movement
1 Roller Free Fixed Free
2 Pin Fixed Fixed Free
3 Rigid Fixed Fixed Fixed
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Reactions
Support reactions for asymmetrical loads and/or supports are computed using the
equations of statics, ΣH=0, ΣV=0, and ΣM=0. The following examples illustrate the use
of the three equations to find reactions.
1 Weight lifter on balcony
Assume:
P = 400#, L = 6’
ΣV=0↑+
R–P=0
R=P R = 400 #
Σ M = 0, +
PL–M=0
2.
M = P L = 400 x 6
Flag pole
- 2 0 0 6 M =2,400 #’
Assume:
9
H = 80# (wind load on flag)
1 9 0
erle
L = 20’
ΣH=0→+
c h i W–H=0
GS
H=W H = 80 #
ΣM=0+
i g h t G WL+M=0
M = - W L = - 80 x 20 M = -1,600 #’
y r Note:
Cop
The negative moment implies, the positive moment arrow must be reversed
3 Tow truck
Assume:
P = 2k, C = 7’, L = 10’
Σ Ma = 0 +
Rb L - P C = 0
Rb = P C / L = 2 x 7 / 10 Rb = 1.4k
Σ Mb = 0 +
Ra L - P (C+L) = 0
Ra = P (C+L) /L = 2 (7+10) / 10 Ra = +3.4k
a b Check Σ V = 0
Σ V = 0 = +3.4 –1.4 –2 ΣV = 0
Note:
The lever arm is always perpendicular to load
Ra Rb Rb pointing downward is provided by the truck weight

Beam reactions
To find reactions for asymmetrical beams:
• Draw an abstract beam diagram to illustrate computations
• Use Σ M = 0 at one support to find reaction at other support
• Verify results for vertical equilibrium
1 Floor framing
2 Abstract beam diagram
Assume:
DL = 10 psf
LL = 20 psf
Σ = 30 psf
Uniform beam load:

w = 30 psf x 2’
- 2 0 0 6 w = 60 plf
0
For convenience, substitute total beam load W for uniform load w at its centroid
9
Total beam load
1 9
W = w L = 60 (12+4) W = 960 #
i erle
Support reactions:
GS c hΣ M b = 0 +
12 Ra – 4 W = 0
Ra = 4 x 960 / 12 Ra = 320 #
i g h t G Σ Ma = 0 +
8 W – 12 Rb= 0
y r 12 Rb = 8 x 960
Cop
Rb = 8 x 960 / 12 Rb = 640 #
Check Σ V = 0 ↑+
W=960
Ra + Rb – W= 320 + 640 – 960 = 0 ΣV=0
Alternate method (use uniform load directly)

Support reactions:
Σ M b = 0 +
12 Ra – 4 x 60 plf x 16’ = 0
Ra = 4 x 60 x 16 / 12 Ra = 320 #
Σ M a = 0 +
8 x 60 x 16 – 12 Rb= 0
12 Rb = 8 x 60 x 16
Rb = 8 x 60 x 16 / 12 Rb = 640 #
Check Σ V = 0 ↑+
Ra + Rb – W= 320 + 640 – 960 = 0 ΣV=0

1 Simple beam with point loads
Assume: P = 1.2k
R = 5 P / 2 = 5 x 1.2 / 2 R=3k
2 Beam with overhang and point loads
- 2 0 0 6
Assume: P = 2k
1
ΣMb=0+
9 9 0
erle
12 Ra - 2x16 - 2x12 - 2x8 - 2x4 = 0
i
Ra = (32 + 24 + 16 + 8) / 12 Ra = 6.67 k
GS c h Σ Ma = 0 +
-12 Rb - 2x4 + 2x4 + 2x8 + 2x12 = 0
Rb = (2 x 8 + 2 x 12) / 12 Rb = 3.33 k
h t G Check Σ V = 0 ↑+
rig
6.67 + 3.33 – 5 x 2 ΣV=0
C o p y 3 Beam with uniform load and point load (wall)

Assume: w = 100 plf, P = 800#
Σ M c = 0 +
16 Ra – 100x16x8 – 800x12 = 0
Ra = (100x16x8 + 800x12) / 16 Ra = 1,400 #
Σ Ma = 0 +
-16 Rc + 100x16x8 + 800 x 4 +800 x16 = 0
Rc = (100x16x8 + 800x4 +800 x16) / 16 Rc = 1800 #
Check Σ V = 0 ↑ +
1400+1800-100x16-800-800 ΣV=0

1 Beam with overhang
Assume: w = 300 plf, C = 3’, L = 15’
X1 = 6, X2 = 9’
Σ Mc = 0+ Σ M b = 0+
Check Σ V = 0 ↑+
ΣV=0
2 Beam with two overhangs

Assume: w = 200 plf. C1 = 5’, C2 = 3’, L = 12
X1 = 5’, X2 = 7’
Σ M c = 0+
12 Rb – 200 x 20 x 7 = 0
6
Rb = 200 x 20 x 7 / 12 Rb = 2333 #
Σ M b = 0+
200 x 20 x 5 - 12Rc –0
0 - 2 0 0
9
Rc = 200 x 20 x 5 / 12 Rc = 1667 #
rle 1 9
Check Σ V = 0 ↑+
2333 + 1667 – 200 x 20 ΣV=0
S c h3
i e Twin beams (treat as 2 beams, due to separation pin joint at b)
Simple left beam: w1 = 100 plf, L1 = 10’
h t G G Ra = Rb = 100 x 10 /2
Right beam: w2 = 150 plf, C1 = 8’, L2 = 20’
Ra = Rb = 500 #
g
X1 = 6’, X2 = 14’, Pb = Rb = 500 #
y r i Σ M d = 0+
Cop
20 Rc – 150 x 28 x 14 –500 x 28 = 0
Rc = (150 x 28x 14 + 500 x 28) / 20 Rc = 3640 #
Σ M c = 0+
150 x 28 x 6 – 500 x 8 – 20 Rd = 0
Rd = (150 x 28 x 6 – 500 x 8) / 20 Rd = 1060 #
Check Σ V = 0 ↑+
3640 + 1060 – 150 x 28 - 500 ΣV=0

Static Determinacy
Static determinacy has theoretical and practical implications. It determines if a structure
can be analyzed by statics or requires another theory. Practically, it defines the degree
of redundancy. Increased redundancy provides added safety. Static determinacy is
defined by one of three conditions:
• Unstable: are unstable and must be avoided
• Determinate: have less redundancy but can be analyzed by statics
• Indeterminate: have most redundancy but cannot be analyzed by statics
Given the practical and theoretical implications it is critical to determine the static
determinacy for any structure concept before proceeding with design. The following
means define static determinacy for beams, trusses, and frames.
6
Beam determinacy
beams is defined as follows:
0 - 2 0 0
Given the three equations of statics, ΣH = 0, ΣV = 0, and ΣM = 0, static determinacy for
1 9
Unstable R<E
9
erle
• Determinate: R=E
• Indeterminate: R>E
S c h i E = number of Equations (3)

R = number of Reactions
ht G G A Unstable beams
i g
B Determinate beams
opy r C
D
Indeterminate beams
Roller support
C E
F
G
Note:
Alternate roller support
Pin support
Rigid support (moment resisting)
Roller supports have one unknown reaction (vertical)

Pin supports have two unknown reactions (horizontal and vertical)
Rigid support has three unknown reactions (horizontal, vertical, rotational)
The degree of indeterminacy (a measure of redundancy) is computed as:
R – E = degree of indeterminacy

Truss determinacy
A Unstable trusses
B Determinate trusses
C Indeterminate trusses
1 External determinacy (support reactions)

2 Internal determinacy (bar forces)
External determinacy is defined as for beams described above. For internal determinacy
- 2 0 0 6
the moment equation, ΣM=0, cannot be used since trusses have pin joints to be
statically determinate. Internal determinacy is defined as follows:
9 9 0
Each bar represents one unknown reaction and each joint has two equations for analysis,
ΣH=0, ΣV=0. The moment equation ΣM=0 cannot be used since determinate trusses
1
erle
have pin joints. Thus internal determinacy is defined as:
i
• Unstable: B+R<2J
GS c h•
•
Determinate:
Indeterminate:
B+R=2J
B+R>2J
i g h t G B = number bars
J = number of joints
y r R = number external reactions
Cop Note:
The degree of indeterminacy is computed as:
B + R – 2 J = degree of indeterminacy

Frame determinacy
Static determinacy for frames is more complex than for beams or trusses and there is no
simple formula to define it. However an intuitive process may be used, starting with
external determinacy as follows:
1 A frame supported by pin and roller is externally determinate like a beam but
internally unstable if internal joints are pins
5 Making the internal joints moment resistant makes the frame determinate
9 Removing a degree of freedom makes a determinate frame indeterminate
A similar process may be applied to multi-story frames as follows:
• One rigid joint at every second story stabilizes adjacent joints and makes the frame
determinate
•
0 6
Additional rigid joints makes a determinate frame indeterminate
- 2 0
0
A Unstable frames
B
9
Determinate frames
1 9
erle
C Indeterminate frames
D Roller support
c h
E
i Alternate roller support
GS
F Pin support
G Fixed support (moment resistant)
G
H Pin joint
r i g h t I Rigid joint (moment resistant)
Copy

Vector Analysis
First used by Leonardo da Vinci, graphic vector analysis is a powerful method to analyze
and visualize the flow of forces through a structure. However, the use of this method is
restricted to statically determinate systems. In addition to forces, vectors may represent
displacement, velocity, etc. Though only two-dimensional forces are described here,
vectors may represent forces in three-dimensional space as well. Vectors are defined by
magnitude, line of action, and direction, represented by a straight line with an arrow and
defined as follows:
Magnitude is the vector length in a force scale, like 1” =10 k or 1 cm=50 kN
Line of Action is the vector slope and location in space
Direction is defined by an arrow pointing in the direction of action
1
- 2 0 0 6
Two force vectors P1 and P2 acting on a body pull in a certain direction. The
resultant R is a force with the same results as P1 and P2 combined, pulling in the
9 0
same general direction. The resultant is found by drawing a force parallelogram [A]
9
or a force triangle [B]. Lines in the vector triangle must be parallel to corresponding
1
erle
lines in the vector plan [A]. The line of action of the resultant is at the intersection
of P1 / P2 in the vector plan [A]. Since most structures must be at rest it is more
c h i useful to find the equilibriant E that puts a set of forces in equilibrium [C]. The
GS
equilibriant is equal in magnitude but opposite in direction to the resultant. The
equilibriant closes a force triangle with all vectors connected head-to-tail. The line
G
of action of the equilibriant is also at the intersection of P1/P2 in the vector plan [A].
r i g h t 2 The equilibriant of three forces [D] is found, combining interim resultant R1-2 of
forces P1 and P2 with P3 [E]. This process may be repeated for any number of
Copy forces. The interim resultants help to clarify the process but are not required [F].
The line of action of the equilibriant is located at the intersection of all forces in the
vector plan [D]. Finding the equilibriant for any number of forces may be stated as
follows:
The equilibriant closes a force polygon with all forces connected head-to-tail,
and puts them in equilibrium in the force plan.
3 The equilibriant of forces without a common cross-point [G] is found in stages:
First the interim resultant R1-2 of P1 and P2 is found [H] and located at the
intersection of P1/P2 in the vector plan [G]. P3 is then combined with R1-2 to find
the equilibriant for all three forces, located at the intersection of R1-2 with P3 in the
vector plan. The process is repeated for any number of forces.

Vector components
Vector components have the same effect on a body as the initial vector. Thus
components relate to a vector as two vectors relate to a resultant or equilibriant.
1 The component forces C1 and C2 in two cables supporting a load P are found by
drawing a force triangle [B] with corresponding lines parallel to the those in the
vector plan [A].
2 Forces in more than two cables supporting a load P are indeterminate [C] and
cannot be found by graphic vector method since there is infinite number of
solutions [D]. A problem with more than two unknown force components requires
consideration of relative cable stiffness (cross-section area, length, and stiffness).
6
Hence we can state:
0 - 0 0
Only two components can be found by graphic vector method
2
This example demonstrates graphic vector analysis: Forces are drawn on a vector
9 9
plan with line of action and direction [E]. The magnitude may be written on each
1
vector or the vector may be drawn at a force scale. A force polygon [F] is drawn
erle
next at a force scale, such as 1” = 1k. For good accuracy, the force scale should
i
be as large as space permits. The line of action of the equilibriant (or resultant) is
GS c h then transposed into the vector plan at the intersection of all force vectors [E].
i g h t G
y r
Cop

Truss Analysis
Graphic truss analysis (Bow’s Notation) is a method to find bar forces using graphic
vectors as in the following steps:
A Draw a truss scaled as large as possible (1) and compute the reactions as for
beams (by moment method for asymmetrical trusses).
B Letter the spaces between loads, reactions, and truss bars. Name bars by
adjacent letters: bar BH between B and H, etc.
C Draw a force polygon for external loads and reactions in a force scale, such as
1”=10 pounds (2). Use a large scale for accuracy. A closed polygon with head-to-
tail arrows implies equilibrium. Offset the reactions to the right for clarity.
D
2 0 0 6
Draw polygons for each joint to find forces in connected bars. Closed polygons
with head-to-tail arrows are in equilibrium. Start with left joint ABHG. Draw a
-
0
vector parallel to bar BH through B in the polygon. H is along BH. Draw a vector
9 9
parallel to bar HG through G to find H at intersection BH-HG.
1
erle
E Measure the bar forces as vector length in the polygon.
c h
F
i Find bar tension and compression. Start with direction of load AB and follow
polygon ABHGA with head-to-tail arrows. Transpose arrows to respective bars in
GS
the truss next to the joint. Arrows pushing toward the joint are in compression;
G
arrows pulling away are in tension. Since the arrows reverse for adjacent joints,
r i g h t G
draw them only on the truss but not on the polygon.
Draw equilibrium arrows on opposite bar ends; then proceed to the next joint with
Copy 1
two unknown bar forces or less (3). Draw polygons for all joints (4), starting with
known loads or bars (for symmetrical trusses half analysis is needed).
Truss diagram
2 Force polygon for loads, reactions, and the first joint polygon
3 Truss with completed tension and compression arrows
4 Completed force polygon for left half of truss
5 Tabulated bar forces (- implies compression)

Truss Example
Some trusses have bars with zero force under certain loads. The example here has zero
force in bars HG, LM, and PG under the given load. Under asymmetrical loads these
bars would not be zero and, therefore, cannot be eliminated. Bars with zero force have
vectors of zero length in the equilibrium polygon and, therefore, have both letters at the
same location.
Tension and compression in truss bars can be visually verified by deformed shape (4),
exaggerated for clarity. Bars in tension will elongate; bars in compression will shorten.
In the truss illustrated the top chord is in compression; the bottom chord is in tension;
inward sloping diagonal bars in tension; outward sloping diagonal bars in compression.
Since diagonal bars are the longest and, therefore, more likely subject to buckling, they
are best oriented as tension bars.
1 Truss diagram
- 2 0 0 6
0
2 Force polygon
3
9 9
Tabulated bar forces (+ implies tension, - compression)
1
erle
4 Deformed truss to visualize tension and compression bars
i
A Bar elongation causes tension
GS c hB Bar shortening causes compression
i g h t G
y r
Cop

Funicular
Graphic vector are powerful means to design funicular structures, like arches and
suspension roofs; providing both form and forces under uniform and random loads.
Arch
Assume:
Arch span L = 150, arch spacing e = 20’
DL = 14 psf
LL = 16 psf
Σ = 30 psf
Uniform load
w = 30 psf x 20’ / 1000 w = 0.6 klf
Vertical reactions
R = w L /2 = 0.6 x 150 / 2
-2 0 0 6 R = 45 k
0
Draw vector polygon, starting with vertical reaction R
Horizontal reaction
e 1 9 9
Maximum arch force (diagonal vector parallel to arch tangent)
H = 56 k
F = 72 k
h
1
i e r l
Arch structure
c
2 Parabolic arch by graphic method
G G S Process:
Draw AB and AC (tangents of arch at supports)
t
Divide tangents AB and AC into equal segments
y r i g h 3
Lines connecting AB to AC define parabolic arch envelope
Arch profile
Cop
Process:
Define desired arch rise D (usually D = L/5)
Define point A at 2D above supports
AB and AC are tangents of parabolic arch at supports
Compute vertical reactions R = w L /2
4 Equilibrium vector polygon at supports (force scale: 1” = 50 k)
Process:
Draw vertical vector (reaction R)
Complete vector polygon (diagonal vector parallel to tangent)
Measure vectors (H = horizontal reaction, F = max. arch force)
Note:
The arch force varies from minimum at crown (equal to horizontal reaction), gradually
increasing with arch slope, to maximum at the supports.

Suspension roof
Assume:
Span L = 300, cable spacing e = 10’, sag f = 30’, height difference h = 50’
DL = 14 psf
LL = 16 psf
Σ = 30 psf
Uniform load
w = 30 psf x 10’ / 1000 w = 0.3 klf
Total load
W = w L = 0.3 x 300 R = 90 k
Draw vector polygon, starting with total load W
Horizontal reaction H = 113 k
6
Vertical reactions
Left reactions
Right reaction
0 - 2 0 0 Rl = 26 k
Rr = 64 k
9
Cable tension
1 9
At left support Tl =115 k
erle
At right support (maximum) Tr = 129 k
c h1
i Cable roof structure
GS
2 Parabolic cable by graphic method
Process:
i g h t G Draw AB and AC (tangents of cable at supports)

Divide tangents AB and AC into equal segments
r
Lines connecting AB to AC define parabolic cable envelop
Copy 3 Cable profile

Process:
Define desired cable sag f (usually f = L/10)
Define point A at 2f below midpoint of line BC
AB and AC are tangents of parabolic cable at supports
Compute total load W = w L
4 Equilibrium vector polygon at supports (force scale: 1” = 50 k)
Process:
Draw vertical vector (total load W)
Draw equilibrium polygon W-Tl-Tr
Draw equilibrium polygons at left support Tl-H-Rl
Draw equilibrium polygons at right support Tr-Rr-H
Measure vectors H, Rl, Rr at force scale
Note:
This powerful method finds five unknowns: H, Rl, Rr. Tl. Tr
The maximum cable force is at the highest support

Random load funiculars
To find funicular form and forces under random load is similar to finding reactions as
described above, with one major difference, it requires a two step approach. In step one
a polygon is drawn using an arbitrary pole. In step two the polygon is redrawn after
finding the correct pole location as intersection of two defining vectors. This method may
be used do find form and forces for suspension cables and arches under various load
conditions.
Suspension cable with random load

1 Vector plan of loads
2 Trial polygon of arbitrary pole F’
3 Vector plan based on arbitrary pole polygon
6
4 Arbitrary trial polygon with real pole at intersection of a and b
5
6
Corrected vector plan
Corrected vector polygon
0 - 2 0 0
9
Process:
1 9
Draw vectors AB, BC, etc. for all loads in trial polygon
erle
Select an arbitrary pole F’ in polygon
Draw polar vectors AF’, BF’, etc. for all loads in polygon
c h i Draw parallel polar vectors at intersection of respective load in plan
GS
Transpose trial closing line a’ from [plan to polygon to find q1
Transpose trial closing line b’ from plan to polygon to find q2
i g h t G Define desired locations of right support in plan

Define desired locations of cable sag at intersection of any load in plan
r
Transpose closing line a between supports from plan to q1 in polygon
Copy Transpose closing line b of sag from plan to q2 in polygon

Intersection of closing lines a and b in polygon define correct pole F
Draw correct polar vectors AF, Bf, etc. for all loads in polygon
Draw parallel polar vectors at intersection of respective load in plan
The corrected vectors will intersect with closing lines a and b in plan
Complete support equilibrium in polygon:
Left support: AF- R1-R1h
Right support: DF- R2h-R2
Measure vector lengths in force scale to complete the process
Note:
The process is based on equilibrium at both supports and intersections of all loads with
the cable.

Arch with random load
The process for arches is similar to cables described above, but with forces reversed
from tension to compression and the polygon pole on the opposite side
Top: vector plan

Bottom: vector polygon
Process:
Draw vectors AB, BC, etc. for all loads in trial polygon
Select an arbitrary pole G’ in the polygon
Draw polar vectors AG’, BG’, etc. for all loads in polygon
Draw parallel polar vectors at intersection of respective loads in plan
Transpose trial closing line a’ from [plan to polygon to find q1
6
Transpose trial closing line b’ from plan to polygon to find q2
Define desired locations of right support
- 2 0 0
Define desired locations of arch rise at intersection of any load
0
9
Transpose closing line a between supports from plan to q1 in polygon
1 9
Transpose closing line b of arch rise from plan to q2 in polygon
erle
Intersection of closing lines a and b in polygon define correct pole G
Draw correct polar vectors AG, BG, etc. for all loads in polygon
c h i Draw parallel polar vectors at intersection of respective load in plan
GS
The corrected vectors will intersect with closing lines a and b in plan
Complete support equilibrium in polygon:
i g h t G Left support: AG- R1-R1h

Right support: EG- R2h-R2
r
Measure vector lengths in force scale to complete the process
Copy Note:
The process is based on equilibrium at both supports and at all intersections of loads with
the arch.

Vector reactions
Reactions put loads in equilibrium and therefore can be found by graphic vector analysis
by similar method used to find equilibriant. The process is illustrated on three examples.
For convenience vectors are defined in a vector plan as for truss analysis. For example,
the vector between A and B in the plan extends from A to B in the polygon.
1 Vector plan of loads and equilibriant

2 Equilibrium polygon of loads and equilibriant
Process:
Draw vectors A-B, B-C, etc. for all loads in polygon
Select an arbitrary pole O in polygon
Draw polar vectors A-O, B-O, etc. for all loads in polygon
2 0 0
Draw and measure equilibriant E to close polygon
- 6
Draw parallel polar vectors at intersection of respective load in plan
0
Draw equilibriant E at intersection of A-O and D-O in plan
Note:
1 9 9
erle
The equilibriant E puts vectors in equilibrium
S c h
3
i Vector plan for two reactions
G
4 Vector polygon for two reactions
G
Process:
ri g h t Draw vectors A-B, B-C, etc. for all loads in polygon

y
Co p Draw parallel polar vectors at intersection of respective load in plan

Draw closing vector E from R1 to R2 in plan
Closing vector E in polygon defines reactions R1 and R2
Measure Reactions R1 an R2
5 Random load vector plan

6 Random load vector polygon
Process:
Draw vectors A-B, B-C, etc. for all loads in polygon
Draw parallel polar vectors at intersection of respective loads in plan
Draw closing vector F from R1 to R2 in plan
Draw parallel vector F in polygon to define length of R1 and R2
Draw and measure reaction R1, R1, H to close the polygon

5
This chapter introduces the theory and examples of strength, stiffness, and stability
described n the following sections: Force types; force vs. stress; allowable stress; axial
St re ngt h stress; shear stress; principle stress and Mohr’s circle; torsion; strain; Hooke’s law;
Poisson’s ratio; creep, elastic modulus; thermal strain; thermal stress; and stability.
St iffne ss
St a bilit y
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop
5-1 MECHANICS Strength Stiffness Stability

Force types
Forces on structures include tension, compression, shear, bending, and torsion. Their
effects and notations are tabulated below and all but bending and related shear are
described on the following pages. Bending and related shear are more complex and
further described in the next chapter.
Type of forces
Force type Action Symbol Notation
Internal reaction
Tension Elongates +
arrows
Internal reaction
Compression Shortens -
arrows
Shear Sliding force Arrow couple Clockwise couple +
Bending
Elongates one side
shortens other side
Concave and
2 0
convex arcs
- 0 6 Concave arc +
Convex arc -
0
Torsion Twists Bar with arrows Right-hand-rule +
1 9 9
erle
1 Axial force (tension and compression)
2 Shear
c h
3
i Bending
GS
4 Torsion
5 Force actions
G
6 Symbols and notations
rig h t A Tension
C o p y B
C
D
E
Compression
Shear
Bending
Torsion

Force vs. stress
Force and stress refer to the same phenomena, but with different meanings. Force is an
external action, measured in absolute units: # (pound), k (kip); or SI units: N (Newton),
kN (kilo Newton). Stress is an internal reaction in relative units (force/area), measured in
psi (pound per square inch), ksi (kip per square inch); or SI units: Pa (Pascal), kPa (kilo
Pascal). Axial stress is computed as:
f=P/A
where
f = stress
P = force
A = cross section area
Note: stress can be compared to allowable stress of a given material.
•
•
Force is the load or action on a member
2 0 0 6
Stress can be compared to allowable stress for any material, expressed as:
-
0
F≥f (Allowable stress must be equal or greater than actual stress)
where
1 9 9
erle
F = allowable stress
f = actual stress
c h i
The type of stress is usually defined by subscript:
GS
Fa, fa (axial stress, capital F = allowable stress)
Fb, fb (bending stress, capital F = allowable stress)
h t G Fv, fv (shear stress, capital F = allowable stress)
rig
The following examples of axial stress demonstrate force and stress relations:
C o p y 1 Wood column (compression)

Assume: Force P = 2000#, allowable stress F = 1000 psi
A = 2 x 2 = 4 in2 (cross section area)
Stress f = P / A = 2000# / 4 f = 500 psi
1000 > 500, ok
2 Steel rod (tension)
Assume: P = 6 k, ½” rod, Fa = 30 ksi
Cross section area A = π r2 = (0.5/2)2 π A = 0.2 in2
Stress f = P / A = 5 k / 0.2 f = 25 ksi
25 < 30, ok
3 Spiked heel on wood stair (compression)
Assume: P = 200# (impact load), A = 0.04 in2, Fa= 400 psi
Stress f = P / A = 200 / 0.04 f = 5000 psi
5000 >> 400. NOT ok
Note: The heel would sink into the wood, yield it and mark an indentation

Wood
Allowable stress Base values for Douglas Fir-Larch 2"-4" (5-10 cm) thick, 2" (5 cm) or wider for allowable
Allowable stress is defined by a material’s ultimate strength or yield strength and a stress: bending (Fb), tension (Ft), compression (Fc), compression normal to grain (FC⊥),
factor of safety. Building codes and trade associations provide allowable stress for horizontal shear (Fv), and elastic modulus (E).
various materials and grades of materials, which may also depend on duration of load. Grade Fb Ft Fc Fc⊥ Fv E units
Allowable wood stress also depends on temperature, moisture content, size, and if a Select 1,450 1000 1,700 625 95 1,900,000 psi
member is single or repetitive, like closely spaced joists. Relevant factors regarding Structural: 9,998 6,895 11,722 4,309 655 13,100,500 kPa
allowable stress are briefly introduced here and further described later in this chapter. No. 1: 1,000 675 1,450 625 95 1,700,000 psi
6,895 4,654 9,998 4,309 655 11,721,500 kPa
Ultimate strength is the stress at which a test specimen breaks under load. Ultimate No. 2: 875 575 1300 625 95 1,600,000 psi
strength varies by material, such as wood, steel, masonry, or concrete, as well as grades 6,033 3,965 8,964 4,309 655 11,032,000 kPa
of each material. Steel
Table of yield stress (Fy); ultimate strength (Fu); allowable stress for bending (Fb),
Yield strength is the point where a material under load changes from elastic to plastic
6
compression (Fc), tension (Ft), shear (Fv); and elastic modulus (E)
0
deformation. Elastic deformation allows the material to return to its unstressed length
0
Steel grade Fy Fu Fb Fc Ft Fv E ksi
2
after the load is removed; by contrast plastic deformation is permanent.
-
ASTM A36 36 58-80 22 22 14.5 29,000 ksi
Factor of safety accounts for uncertainty regarding consistency of material quality, type
of stress (tension, compression, shear, bending) and actual load conditions. The factor
ASTM A572
1 9
248
50
400-550
65
9
150
30
0 150
30
100
20
200,000
29,000
MPa
ksi
erle
345 450 210 210 140 200,000 MPa
of safety is defined differently for different materials. For example, for steel the factor of Masonry
i
safety is based on yield strength, for concrete on the specified compressive strength
h
Allowable compressive stress Fa, for masonry with special inspection is 25% of
c
(breaking strength). The tables at left give some typical allowable stresses.
GS
specified strength ƒ’m by the working stress method; reduced for slenderness. Specified
Compressive strength ƒ’m is based on compressive strength of masonry units and
G
mortars type M, S, N.
r i g h t Type
Unit strength 1.9
Concrete masonry (ksi)
2.8 3.75 4.8
Clay brick masonry
4 6 8 10 12 14
(ksi)
Cop y ƒ’m (M or S)
ƒ’m (N)
Type
Unit strength
1.5 2
1.35 1.85 2.35
2.5 3
2.8 1.6
2
Concrete masonry (MPa) Clay brick masonry

13 19 26 33 28
2.7 3.35
2.2
41
2.7
55
4
3.3
69
4.7 5.3
3.8 4.4
(MPa)
83 97
ƒ’m (M or S) 10 14 17 21 14 19 23 28 32 37
ƒ’m (N) 9 13 16 19 11 15 19 23 26 30
Concrete
By working stress method, allowable stresses are based on compressive strength ƒ‘c.
Typical compressive strengths range from 2 to 6 ksi (14 to 41 MPa)
Allowable compressive stress 0.40 ƒ‘c
Allowable compressive bending stress 0.45 ƒ‘c
Allowable shear stress without reinforcing: beam 1.1 ƒ‘c1/2
joist 1.2 ƒ‘c1/2
footing & slab on grade 2.0 ƒ‘c1/2
Note: For concrete strength design method see chapter 8.

Axial stress
Axial stress acts in the axis of members, such as columns. Axial tension is common in
rods and cables; wile axial compression is common in walls and columns. The following
examples illustrate axial design and analysis. Analysis determines if an element is ok;
design defines the required size. The equation, fa = P/A, is used for analysis. The
equation A = P/Fa, is used for design. Allowable stress, Fa, includes a factor of safety.
1 Crane cable design
Assume: P = 12 k, Fa = 70 ksi
Find required cable size
Metallic cross section Am (cables are about 60% metallic)
Am = P / Fa = 12 k / 70 ksi Am = 0.17 in2
Gross cable area
Ag = Am/ 0.6 = 0.17 / 0.6 Ag = 0.28 in2
Cable size
φ = 2 (A/π)1/2 = 2 (0.28 / π)1/2 = 0.6”
- 2 0 0 6 use φ = 5/8”
0
2 Suspension hanger analysis (Hong Kong-Shanghai bank)
9 9
Assume: load per floor P=227 k, Fa=30 ksi, level 1 A = 12 in2. level 6 A = 75 in2
1
erle
Hanger stress
Level 1: fa = P / A = 227 / 12 fa = 19 ksi < 30
c h i Level 6: fa = 6 P / A = 6 x 227 / 75 fa = 18 ksi < 30
GS
3 Post/footing analysis
Assume: P = 12,000 #, 3’x3’x2’ footing at 150 pcf, 4x4 post (3.5”x3.5” actual)
G
Allowable post stress Fa = 1000 psi, allowable soil pressure Fs = 2000 psf
r i g h t Post stress
P/A = 12,000 # / (3.5”x 3.5”) fa = 980 psi < 1000
Copy 4
Soil pressure
fs = P/A = (12,000 # + 3’x3’x2’x150 pcf)/ (3’x3’)
Slab/wall/footing, analyze a 1’ wide strip
fs = 1633 psf < 2000
Assume: allowable wall stress Fa = 360 psi; allowable soil pressure Fs=1500 psf
Concrete slab, t =8” thick, L = 20’ span
CMU wall, h =10’, DL = 80 psf, t = 8” nominal (7 5/8” = 7.625” actual)
Slab load
100 psf DL+ 40 psf LL DL+LL =140 psf
Load at wall base
P = 140psf (20’/2) + 80 psf (10’) P = 2,200 #
Wall stress
fa = P / A = 2200 # / (12”x7.625”) fa = 24 psi < 360
Load on soil
P = 2200 + 150 pcf x 2’ x 1’ P = 2,500 #
Soil pressure
fs = 2,500 # / (1’ x2’) fs = 1250 psf < 1500

Shear stress
Shear stress occurs in many situations, including the following examples, but also in
conjunction with bending, described in the next chapter on bending. Shear stress
develops as a resistance to sliding of adjacent parts or fibers, as shown on the following
examples. Depending on the number of shear planes (the joining surface [A] of
connected elements) shear is defined as single shear or double shear.
A Shear plane
B Shear crack
1 Single shear
Assume: P = 3 k = 3000 #, 2”x4” wood bars with ½” bolt of Fv = 20 ksi
Shear area (bolt cross section)
A = π r2 = π (0.5/2)2 A = 0.2 in2
Shear stress fv = P / A = 3/ 0.2
- 2 0 0 6 fv = 15 ksi < 20
0
2 Check end block (A)
1 9
End block shear area A = 2x 2” x 6”
9
Assume: Block length 6”, wood Fv = 95 psi, all other as above
A =24 in2
erle
Shear stress fv = P / A = 3000# / 24 fv =125 psi > 95
i
NOT ok
G S c h Required block length e = 125 x 6”/ 95 = 7.9: use e = 8
G
3 Double shear
t
Assume: P = 22 k, 2 5/8” bolts of Fv = 20 ksi
y ri g h Shear area A = 4π r2 = 4 π (0.625/2)2

Shear stress fv = P / A = 22 / 1.2
A = 1.2 in2
fv = 18 ksi < 20
Co p 4 Double shear, glued

Assume: P = 6000 #, Wood bars, Fv = 95 psi
Shear area A = 2 x 4” x 8” A = 64 in2
Shear stress Fv = P / A = 6000 / 64 fv =94 psi < 95
5 Twin beam double shear

Assume: P = R = 12 k, 2 ½“ bolts, Fv = 20 ksi
Shear area A = 4π r2 = 4 π (0.5/2)2 A = 0.79 in2
Shear stress fv = P / A = 12 / 0.79 fv = 15 ksi < 20
6 Shear wall
Assume: P = 20 k, 8” CMU wall, t = 7.625”, L = 8’, Fv = 30 psi
Shear area A = 7.625” x 12” x 8 ‘ A =732 in2
Shear stress fv = P / A = 20,000 # / 732 fv = 27 psi < 30

Shear cracks
Shear cracks are diagonal, even for horizontal seismic forces. Although intuitive
assumption suggest horizontal cracks for lateral load. The diagonal cracks are described,
considering a square wall element subject to lateral load.
1 Lateral load generates a horizontal reaction at the base

The horizontal force couple tends to rotate the wall clockwise
A counterclockwise couple provides rotational equilibrium, Σ M = 0.
2 Incorrect intuitive assumption of horizontal shear crack
3 The horizontal and vertical couples presented as two vectors each
6
4 Combined shear vectors at each corner yield diagonal vectors
- 2 0 0
The diagonal vectors yield compression in one direction
The diagonal vectors yield tension in the opposite direction
0
9
The tension generates cracks
1 9
erle
5 Reversed earthquake shaking generates cracks in opposite direction
c h6
i Typical X-cracks caused by earthquakes
h t G GS
y r i g
Cop
X-cracks caused by Northridge Earthquake

Torsion
Torsion is very common in machines but less common in building structures. The
examples here include a small detail and an entire garage.
1 Door handle
Assume: P = 10 #, e = 3”
Torsion moment M
M = P e = 10 x 3 M = 30 #’
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
rig
2 Tuck-under parking
y
Assume: Shear e =10’, base shear V = 12 k
C o p Torsion moment M
M = V e = 12 k x 10’ M = 120 k’
Note:
The torsion moment is the product of base shear v and lever arm e, the distance from
center of mass to center of resistance (rear shear wall).
In the past, torsion of tuck-under parking was assumed to be resisted by cross shear
walls. However, since the Northridge Earthquake of 1994 where several buildings with
tuck-under parking collapsed, such buildings are designed with moment resistant
beam/column joints at the open rear side.

Principle stress
Shear stress in one direction, at 45 degrees acts as tensile and compressive stress,
defined as principle stress. Shear stress is zero in the direction of principle stress, where
the normal stress is maximum. At any direction between maximum principle stress and
maximum shear stress, there is a combination of shear stress and normal stress. The
magnitude of shear and principle stress is sometimes required for design of details.
Professor Otto Mohr of Dresden University develop 1895 a graphic method to define the
relationships between shear stress and principle stress, named Mohr’s Circle. Mohr’s
circle is derived in books on mechanics (Popov, 1968).
Isostatic lines
Isostatic lines define the directions of principal stress to visualize the stress trajectories in
6
beams and other elements. Isostatic lines can be defined by experimentally by photo-
0 - 2 0
elastic model simulation or graphically by Mohr’s circle.
0
9
1 Simple beam with a square marked for investigation
9
2 Free-body of square marked on beam with shear stress arrows
3
4
erle 1
Free-body square with shear arrows divided into pairs of equal effect
Free-body square with principal stress arrows (resultant shear stress vectors)
c h
5
i Free-body square rotated 45 degrees in direction of principal stress
GS
6 Beam with isostatic lines (thick compression lines and thin tension lines)
G
Note:
r i g h t Under gravity load beam shear increases from zero at mid-span to maximum at supports.
Beam compression and tension, caused by bending stress, increase from zero at both
Copy supports to maximum at mid-span. The isostatic lines reflect this stress pattern; vertical
orientation dominated by shear at both supports and horizontal orientation dominated by
normal stress at mid-span. Isostaic lines appear as approximate tension “cables” and
compression “arches”.

Strain
Strain is a deformation caused by stress, or change in temperature, described later.
Strain may elongate or shorten a solid, depending on the type of stress.
Hooke’s law
Material expands and contracts under tension and compression, respectively. The
stress/strain relationship, called Hooke’s law after the English scientist Robert Hooke,
who discovered it in the 17th century, has since been confirmed by many empirical tests.
The Hooke’s law assumes isotropic material (equal properties in any direction). The
stress/strain relation is visualized here by a spring, as substitute for rods as used in
testing machines, to amplify the deformation.
1 Elongation due to tension
6
2 Shortening due to compression
3
L
Stress / strain graph
Unstressed length
0 - 2 0 0
9
ΔL Strain (elongation or shortening under load)
P
1 9
Applied load
erle
ε Unit strain Epsilon (ε = ΔL/L)
Elastic modulus E = f / εf
i
E Stress f = P / A
GS c hA Cross section area of assumed rod
i g h t G
y r
Cop

Stress/strain relations
Although stress/strain tests may be done for any materials, for convenience the following
test description assumes a steel rod. After measuring the unstressed length, load is
applied and the strain recorded. The load is then incrementally increased and all related
elongations recorded on a Cartesian graph, strain on the horizontal axis, and stress on
the vertical axis. The recorded measure points are connected by a line. A straight line
implies linear stress/strain relations; a curved line implies non-linear relations. Most
structural materials are linear up to the proportional limit, and non-linear beyond that
point. If the rod returns to its original length after the load is removed, the material is
considered elastic; if it remains deformed it is considered plastic. The remaining
deformation is the permanent set. Rubber is an elastic material; clay a plastic material.
Some materials, such as steel, are elastic-plastic, i.e., up to the elastic limit steel is
elastic; beyond the elastic limit it is plastic. The transition from elastic to plastic strain is
6
also called yield point. Materials which deform much and absorb energy before breaking
- 2 0 0
are considered ductile; materials which break abruptly are considered brittle. Mild steel is
considered a ductile material; concrete is usually brittle.
0
1
2
9
Test loads 1 to 5 kip
1 9
Stress-strain graph (horizontal axis = strain, vertical axis = stress)
erle
3 Linear material (linear stress/strain relation
c h
4
5
i Non-linear material (non-linear stress /strain relation)
Elastic material (returns to original size if unloaded, like rubber)
GS
6 Plastic material (remains permanently deformed like clay)
G
7 Brittle material (breaks abruptly)
r i g h t 8 Ductile material (deforms and absorbs energy before breaking)
Elastic modulus
Copy E
f
ε
S
E = f / ε = Elastic Modulus (defines material stiffness)
Stress
Unit strain (ε = ΔL/L)
Permanent set (remaining strain after stress is removed)
Derivation of working equation to compute strain:

ΔL f P
=ε= = solving for ΔL
L E AE
ΔL = PL / AE
The equation is used to compute strain due to load. It shows that strain:
• Increases with increasing P and L
• Increases inversely with A and E

Poisson’s ratio
Poisson’s ratio is named after French scientist Poisson who defined it 1807 as ratio of
lateral strain / axial strain. All materials shrink laterally when elongated and expand when
compressed. Poisson’s ratio is defined as:
ν = lateral strain / axial strain
Based on empirical tests, Poisson’s ratio for most materials is in the range of 0.25 to
0.35; only rubber reaches 0.5, the maximum for isotropic material.
Creep
Creep is a time dependent strain, most critical in concrete where it is caused by moisture
squeezed from pores under stress. Creep tends to diminish with time. Concrete creep
may exceed elastic strain several times, as demonstrated by Case Study 9 of Northridge
Earthquake failures (Schierle, 2002). Yet much research is needed to provide design
data and guidelines regarding creep.
- 2 0 0 6
Elastic modulus
9 9 0
The elastic modulus E, also called modulus of elasticity or Young’s modulus Y, after
1
erle
English scientist Young, who defined it 1807. The term elastic modulus is actually a
misnomer since it defines stiffness, the opposite of elasticity.
c h i
Note:
GS
Since E=f/ε and ε is a ratio without units, the elastic modulus has the same units as
stress
ig h t G 1 Poisson’s ratio effect
r
2 Creep deformation (C = creep, T = time)
Co p y 3
4
5
Elastic / plastic stress / strain curve (E = elastic range, P = plastic range)
Abstract steel graph (A = proportional limit, B = elastic limit, C = yield point,
CD = yield plateau, E = ultimate strength, F = breaking point)
Mild steel stress / strain curve
6 High strength steel stress / strain curve
7 Concrete stress / strain curve (compressive strengths: A=9 ksi, B=4 ksi, C=3 ksi)
8 Stress / strain of linearly elastic wood
Allowable stress vs. elastic modulus (typically about 1:1000 ratio)
Material Allowable stress (psi) Elastic modulus (psi)
Wood 1,400 1,400,000
Steel 30,000 30,000,000
Masonry 1,500 1,500,000
Concrete 3,000 3,000,000

Strain examples
Elevator cables
Assume
4 cables φ ½” each, 60% metallic areaBreaking strength Fy = 210 ksi
Allowable stress (210 ksi / 3) Fa = 70ksi
Elastic Modulus E = 16,000ksi
L = 800’ each
P = 8k
Metallic area
Am = 4 π r2 = 4 x .6 π (0.5/2)2 Am = 0.47in2
Stress
f = P / A = 8 / 0.47 f = 17ksi
17 < 70, ok
6
Elongation under load
ΔL = PL / AE
ΔL = 8k x 800” x 12” / (0.47x16000)
0 - 2 0 0 ΔL = 10”
1 9
Suspended building
9
erle
3 Differential strain
i
Assume
GS c h 10 stories @ 14’ = 10x14’x12”

Average column stress
Average strand stress
L = 1680”
f = 18 ksi
f = 60 ksi
i g h t G Elastic modulus (steel)

Elastic modulus (strand)
E = 29,000 ksi
E = 22,000 ksi
y r ΔL = PL/AE, since f = P/A → ΔL = f L/E
Cop
Column strain
ΔL = 18ksi x 1680” / 29000 ΔL = 1”
Strand strain
ΔL = 60ksi x 1680” / 22000 ΔL = 4.6”
Differential settlement ΔL = 5.6”
4 Shorten hangers under DL to reduce differential strain, or prestress strands to

reduce ΔL by half
Note: Differential strain is additive since both strains are downwards

To limit differential strain, suspended buildings have <= 10 stories / stack

Thermal strain
Unrestrained objects expand and contract if subjected to temperature increase and
decrease, respectively. Thermal strain is defined by a coefficient α for each material.
Thermal strain varies linearly with temperature variation.
1 Bar of initial length L
2 Thermal strain ΔL due to temperature increase, computed as:
ΔL = α Δt L
where
α = thermal coefficient (in/in/oF) [/oC (SI units)]
Δt = temperature increase (+) / decrease (-)
L = initial length
Thermal stress
- 2 0 0 6
Thermal stress is caused when thermal strain is prevented by restrains.
3
4
Bar of initial length L
1 9
Elongation ΔL due to heat
9 0
erle
5 Heated bar reduced to initial length by load P
6 Restrained bar under stress
c h i
Thermal stress derivation:
GS
Since ΔL = PL / AE and f = P/A
ΔL = f L / E → f = E ΔL/L
i g h t G ΔL = α Δt L → f = E α Δt L/L
f = α Δt E
y r
Cop
where
f = thermal stress
E = elastic modulus
Coefficient of thermal expansion α and elastic modulus E

Material US α US Eα SI α SI Eα
(10-6/oF) (106psi) (10-6/oC) (106gPa)
Aluminum 13 10 24 69
Steel 6.5 29 11.7 200
Concrete 6 3–4 11 20 – 28
Masonry 4 1–3 7 7 – 21
Wood 1.7 – 2.5 1.2 – 2.2 3.5 - 4.5 8 – 15
Glass 44 9.6 80 66
Plastics 68 - 80 0.3 – 0.4 122 - 144 2 – 2.8
Aluminum 13 10 24 70

Thermal effects
Thermal variations cause stress and/or strain in structures. Temperature increase and
decrease cause material to expand and contract, in unrestrained objects and stress in
restrained objects, respectively, as illustrated in the examples at left.
1 Wall subject to bending stress due to temperature variation
1 Expanding south column causes bending stress in beams
2 Expanding south column causes tensile stress in bracing
3 Expanding fix-end arch subject to reversed bending stress
4 Expanding pin-joined arch subject to bending stress
5 Three-hinge arch, free to expand, without bending stress
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Thermal examples
1 Curtain wall
Assume:
Aluminum curtain wall, find required expansion joint
Δt = 100 o F (summer vs. winter temperature)
2 story mullion, L = 30’ x 12” = 360”
α = 13 x 10-6 in/in/o F
E = 10 x 10 6 psi
Thermal strain
ΔL = α Δt L = 13 x 10-6 x 100 o x 360” ΔL = 0.47”
Use ½” expansion joints 0.5 > 0.47
Assume ignorant designer forgets expansion joint
Thermal stress:
f = α Δt E = 13x10-6x100x10x10-6 psi
2 0 0
Note: 106 and 10-6 cancel out and can be ignored
- 6 f = 13,000 psi
1 9 9 0
13,000 psi is too much stress for aluminum
High-rise building, differential expansion
erle
Assume:
i
Steel columns exposed to outside temperature
GS c h Δt = 50oF (south vs. north temperature)

L = 840’ (60 stories at 14’)
α = 6.5 x 10-6 in/in/o F
ig h t G Differential expansion
r
ΔL = α Δt L = 6.5 x 10-6 x 50 o x 840’ x 12” ΔL = 3.3”
Co p y 3
Note: the differential expansion would cause bending stress
Masonry expansion joints
(masonry expansion joints should be at maximum L = 100’)
Assume
Temperature variation Δt = 70o F
Joint spacing L=100’ x 12” L = 1200”
Thermal coefficient α = 4x10-6/oF
E-modulus E = 1.5x106 psi
Required joint width
L ΔL = α Δt L = (4 x 10-6) 70o (1200”) ΔL = 0.34”
Use 3/8” expansion joint 0.375 > 0.34
Check thermal stress without expansion joint
f = α Δt E
f = 4x10-6 x 70o x 1.5x106 f = 280psi

Stability
Stability is more complex and in some manifestations more difficult to measure than
strength and stiffness but can be broadly defined as capacity to resist:
• Displacement
1 2 3 • Overturning
• Collapse
• Buckling
Diagrams 1-3 give a theoretical definition; all the other diagrams illustrate stability of
conceptual structures.
1
2
Unstable
Neutral
- 2 0 0 6
0
3 Stable
4
9 9
Weak stability: high center of gravity, narrow base
1
erle
5 Strong stability: low center of gravity, broad base
6 Unstable post and beam portal
c h
7
i Stable moment frame
GS
8 Unstable T-frame with pin joint at base
9 Stable twin T-frames
i g h t G
y r
Cop

Overturn stability
To resist overturning under lateral load requires a stabilizing moment greater than the
overturning moment (usually with a safety factor of 1.5). Stabilizing moments are dead
P P weight times lever arm (distance from center of mass to edge of resisting element,
assuming a rigid body) as demonstrated on the following examples (assuming uniform
wind load for simplicity).
1 Building of vertical extrusion

Assume: 20 stories, 90’x90’, B = 90’, h = 300
Wind force F = 70 psf x 90 x 300 / 1000 F = 1,890 k
Overturning moment Mo = F h/2 = 1890x300/2 Mo = 283,500 k’
Dead load P = 50 psf x 902 x 20/1000 P = 8.100 k
6
Stabilizing moment Ms = P B/2 = 8100 x90/2 Ms = 364,500 k’
Check stability (MS > Mo ?)
0 - 2 0 0 364,500 > 283,500
9
1 Building with cantilever core
rle 1 9
Assume: 20 stories, 90’x90’, B = 300’, h = 300
Wind force F = 70 psf x 90 x 300/1000 F = 1,890 k
e
Overturning moment Mo = F h/2 = 1890x300/2 Mo = 283,500 k’
S c h i Dead load P = 50 psf x 902 x 20/1000

Stabilizing moment Ms = P B/2 = 8100x30/2
P = 8.100 k
Ms = 121,500 k’
G
Check stability (MS > Mo ?) 121,500 < 283,500
i g h t G Core is unstable without tension piles or large footing
r
3 Pyramid
Copy Assume: 9 stories, h = 108’, B = 204’

Dead load
Overturn moment Mo = F h/3 = 750x108/3
Stabilizing moment Ms = P B/2 = 4800x204/2
F = 750 k
P = 4800 k
27,000 k’
Ms = 489,600 k’
Check stability (MS > Mo ?) 489,600 >> 27,000
4 Inverted Pyramid
Assume: 9 stories, h = 108’, B = 12’ F = 750 k
P Dead load P = 4800 k
Overturn moment Mo = F 2/3 h = 750 x 2/3 x 108 54,000 k’
Stabilizing moment Ms = P B / 2 = 4800 x 12 / 2 Ms = 28,800 k’
Check stability (MS > Mo) 28,800 < 54,000
Inverted pyramid is unstable without tension piles or large footing

Buckling stability
Buckling stability is more complex to measure than strength and stiffness and largely
based on empirical test data. This introduction of buckling stability is intended to give
only a qualitative intuitive understanding.
Column buckling is defined as function of slenderness and beam buckling as function of

compactness. A formula for column buckling was first defined in the 18th century by
Swiss mathematician Leonhard Euler. Today column buckling is largely based on
empirical tests which confirmed Euler’s theory for slender columns; though short and
stubby columns may crush due to lack of compressive strength.
Beam buckling is based on empirical test defined by compactness, a quality similar to
6
column slenderness.
- 2 0 0
Slender column buckles in direction of least dimension
0
9
2 Square column resist buckling equally in both directions
3
9
Blocking resists buckling about least dimension
1
erle
4 Long and slender wood joist subject to buckling
5 blocking resists buckling of wood joist
c h
6
i Web buckling of steel beam
GS
7 Stiffener plates resist web buckling
G
A Blocking of wood stud
r i g h t B
C
Blocking of wood joist
Stiffener plate welded to web
Copy P Load

6 Bending elements are very common in structures, most notably as beams. Therefore,
the theory of bending is also referred to as beam theory, not only because beams are the
most common bending elements but their form is most convenient to derive and describe
the theory. For convenience, similar elements, such as joists and girders, are also
considered beams. Although they are different in the order or hierarchy of structures,
their bending behavior is similar to that of beams, so is that of other bending elements,
such as slabs, etc., shown on the next page. Thus, although the following description
applies to the other bending elements, the beam analogy is used for convenience.
Be nding Beams are subject to load that acts usually perpendicular to the long axis but is carried in
bending along the long axis to vertical supports. Under gravity load beams are subject to
bending moments that shorten the top in compression and elongate the bottom in
2 0 0 6
tension. Most beams are also subject to shear, a sliding force, that acts both horizontally
and vertically. Because beams and other bending elements are very common, the beam
-
0
theory is important in structural design and analysis.
1 9 9
As for other structural elements, beam investigation may involve analysis or design;
erle
analysis, if a given beam is defined by architectural or other factors; design, if beam
dimensions must be determined to support applied loads within allowable stress and
c h i
deflection. Both, analysis and design require finding the tributary load, reactions, shear,
GS
and bending moment. In addition, analysis requires to find deflections, shear- and
bending stress, and verify if they meet allowable limits; by contrast design requires sizing
G
the beam, usually starting with an estimated size.
rig h t The following notations are commonly used for bending and shear stress:
y
fb= actual bending stress Fb= allowable bending stress
C o p fv= actual shear stress Fv= allowable shear stress

Allowable stresses are given in building codes for various materials.
Allowable stresses assumed in this chapter are:
Wood
Fb= 1450 psi (9998 kPa) Fv= 95 psi (655 kPa)
Steel
Fb= 22 ksi (152 MPa) Fv= 14 ksi (97 MPa)
The more complex design and analysis of concrete and masonry will be introduces-later.
6-1 FUNDAMENTALS: Bending

Bending elements
As mentioned above, beams are the most common bending elements and their structural
behavior described in this chapter applies in general to other bending elements as well.
Other bending elements, explored later, include joists, girders, slabs and plates
(analyzed as strip of unit width) as well as folded plates, cylindrical shells, moment
frames, and Vierendeel girders (named after the Belgian inventor of the 19th century).
Thus, the theory of bending and shear has broad implications and is very important for
structural analysis and design.
1 Beams, one shown deformed under uniform gravity load
2 Slab or deck with a strip of unit width marked for analysis as “beam”
3 Folded plate acts as narrow, inclined beams leaning against one-another
4 Cylindrical shell, acts as beams of semi-circular cross-section
5
6
2 0 0 6
Moment frame resists gravity and lateral load in bending
Vierendeel girder (named after the Belgian inventor of it)
-
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Beam types
The location, number, and type of supports determine the type of beam.
1 Simple beam
2 Cantilever beam
3 Beam with one overhang
4 Beam with two overhangs
5 Restrained beam
6 Continuous beam
The simple beam is most common in practice. It has two supports, one pin and one
roller, and, with three unknown reactions, it is statically determinate. Given their pin and
roller supports, overhang beams are also determinate; by contrast restraint and
6
continuous beams are statically indeterminate, since they have more than three unknown
reactions.
0 - 2 0 0
The simple beam has single concave curvature that results in compression on top and
9 9
tension at the bottom of the beam. The cantilever beam has single convex curvature,
1
with tension on top and compression at the bottom of the beam. All the other beams
erle
change from concave to convex curvatures. Because the cantilever beam has only one
c h i
support, it must have fixed (moment resistant) support to be stable.
GS
Given equal loads and spans, the cantilever beam has the largest bending moment,
followed by the simple beam with half that of the cantilever. Beams with overhang have
i g h t G negative overhang moments that reduces positive field moments. Their reduced bending
moment is less than the moment of a simple beam of equal span. Two overhangs yield
r
smaller field moments than a single overhang. Designing a beam with one or two
Copy overhangs is a good strategy to greatly reduce bending for better efficiency without extra
cost. Given the double curvature of restrained and continuous beams, they, too, have
reduced bending moments. In restrained beams this advantage may be in part offset by
the fact that the negative end moments must be resisted by supports. The interaction of
beam and column provides lateral resistance for moment frames.

Bending and Shear
Although derivation and numeric examples are required to analyze and design beams,
intuitive understanding is an important prerequisite to gain deeper insight into the
behavior of beams. The following is an intuitive introduction to beam bending and shear.
A simple beam with uniform load is used for convenience.
Bending moment
Gravity load on a simple beam shortens the top and elongates the bottom, causing
compressive and tensile stresses at top and bottom, respectively; with zero stress at the
neutral axis (N. A.). In beams of symmetrical cross-section, the neutral axis is at the
center. The compressive and tensile stress blocks generate an internal force couple that
resists the external bending moment caused by load.
1 Simple beam with pin and roller supports
- 2 0 0 6
0
2 Deformed beam under uniform gravity load
3
9 9
Free-body diagram with bending stress block that generates an internal force
1
couple to resist the external bending moment caused by load
c h i erle
h t G GS
y r i g
Cop

Shear force
With few exceptions, described later, shear coexists with bending. When shear is
present it acts both horizontally and vertically at equal magnitude. In wood beams
horizontal shear is more critical because wood’s shear capacity is much smaller parallel
than perpendicular to the grain.
1 Beam under uniform load with shear cracks as they occur in some concrete beams
near the supports where shear is maximum
2 Tendency of beam parts to slide vertically generates vertical shear stress that is zero
at mid-span and increases to maximum at the supports where the vertical shear
deformation is greatest
3 Tendency of beam layers to slide horizontally generates horizontal shear that is zero
composed of several boards
- 2 0 0 6
at mid-span and increases toward the supports. This is visualized, assuming a beam
0
4 Shear diagram reflects shear distribution over beam length
9 9
5 Unloaded beam marked with squares to visualize shear
1
erle
6 Loaded beam with squares deformed into rhomboids due to shear
7 Horizontal and vertical shear couples on a square beam part are equal to balance
c h irotational tendencies (ΣM = 0). Therefore, horizontal and vertical shear stresses are
GS
equal at any point on the beam.
8 Shear vectors generate compression and tension diagonal to the shear. This tends
G
to generate diagonal tension cracks in concrete beams
r i g h t
Copy

Bending and shear distribution
Shear and bending diagrams illustrate their respective distribution over the beam’s length
(simple beam in this case). Similarly, internal shear and bending stresses, caused by
shear force and bending moment, respectively, may be drawn to illustrate their
distribution over the beam’s cross-section.
1 Beam diagram
2 Shear diagram shows shear force distribution over beam length
3 Bending diagram shows bending moment distribution over beam length
4 Shear stress diagram shows distribution over beam depth
5 Bending stress diagram shows distribution over beam depth
A Possible location of a pipe hole at beam center and mid-span where both shear
6
and bending stresses are zero
2 0 0
The shear force varies linearly from maximum positive shear at the left support to
0 -
maximum negative shear at the right support. The left reaction pushing upward and the
9 9
beam load downward cause a maximum positive (clockwise) shear couple at the left
1
support. Shear reduces to zero at mid-span where it is balanced by gravity load. At the
erle
right support shear reaches a negative maximum.
c h i
The bending moment varies in parabolic form over the beam’s length. It is zero at both
GS
supports and maximum at mid-span. Over the depth of the beam, bending stress varies
from maximum compression on the top to maximum tension at the bottom.
i g h t G The variation of bending over the beam length creates differential bending stress that is
unbalanced. Thus, the compressive and tensile bending stresses push in opposite
y r directions which causes horizontal shear stress. Shear stress varies from zero on top
Cop
and bottom to maximum at the neutral axis. The rare case of uniform bending over the
beam length, i.e., no differential bending stress, causes zero shear stress. This is called
pure bending.

Bending and shear stress
Bending and shear stresses in beams relate to bending moment and shear force similar
to the way axial stress relates to axial force (f = P/A). Bending and shear stresses are
derived here for a rectangular beam of homogeneous material (beam of constant
property). A general derivation follows later with the Flexure Formula.
1 1 Simple wood beam with hatched area and square marked for inquiry
2 Shear diagram with hatched area marked for inquiry
3 Bending moment diagram with hatched area marked for inquiry
4 Partial beam of length x, with stress blocks for bending fb and shear fv,
where x is assumed a differential (very small) length
Reactions, found by equilibrium ΣM = 0 (clockwise +)
at c: +12 Ra-3(8)=0; Ra=3(8)/12 Ra = 2 k
6
2 at a: -12 Rc+3(4)=0; Rc=3(4)/12 Rc = 1 k
right of a and left of b
0
V= 0+2
- 0 0
Shear V, found by vertical equilibrium, ΣV=0 (upward +).
2 V=2k
1 9
right of b and left of c
9
V= 2-3
Bending moment M, found by equilibrium ΣM=0 (clockwise +)
V = -1 k
erle
at a: M = +2(0) M = 0 k’
c h i at b: M = +2(4) M = 8 k’
GS
at c: M = +2(12)-3(8) M = 0 k’
3
Bending stress fb is derived, referring to 4. Bending is resisted by the force couple C-T,
G
with lever arm 2/3 d = distance between centroids of triangular stress blocks. C=T= fb
r i g h t bd/4, M= C(2d/3) = (fbbd/4)(2d/3) = fbbd2/6, or

fb= M/(bd2/6); where bd2/6 = S= Section Modulus for rectangular beam; thus
Cop y fb= M/S

For our beam: S= bd2/6= 4(12)2/6=
fb= M/S= 8(1000)12/96 *
S = 96 in3
fb = 1000 psi
* multiplying by 1000 converts kips to pounds, by 12 converts feet to inches.
Shear stress fv is derived, referring to 4. Bending stress blocks pushing and pulling in
opposite directions create horizontal shear stress. The maximum shear stress is fv=C/bx,
where b = width and x = length of resisting shear plane. Shear at left support is V = R.
Let M = bending moment at x, and fb= bending stress at x, then M = RX = VX, and
fb=M/S=Vx/S. Substituting Vx/S for fb in C = fbbd/4, the compressive top force, yields
C=(Vx/S)(bd/4). Thus fv=C/(bx) yields fv=(Vx/S)(bd/4)/(bx). Substituting bd2/6 for S yields
Vx bd / 4 Vbd / 4 6V
4 fv = 2
= 2 2 = , or
bd / 6 bx b d / 6 4bd
fv = 1.5 V / bd
For the sample beam: fv = 1.5(2)1000/(4x12) fv = 63 psi

Equilibrium Method
Cantilever beam with point load
Assume a beam of length L = 10 ft, supporting a load P = 2 k. The beam bending
moment and shear force may be computed, like the external reactions, by equations of
equilibrium ΣH=0, ΣV= 0, and ΣM=0. Bending moment and shear force cause bending-
and shear stress, similar to axial load yielding axial stress f= P/A. Formulas for bending-
and shear stress are given on the next page and derived later in this chapter.
1 Cantilever beam with concentrated load
6
V Shear diagram (shear force at any point along beam)
M
Δ
- 2
Deflection diagram (exaggerated for clarity)
0 0 0
Bending moment diagram (bending moment at any point along beam)
9 9
Reactions, found by equilibrium, ΣV=0 (up +) and ΣM=0 (clockwise +)
1
rle
at b ΣV= 0 = R–2= 0 R=2k
e
at b ΣM= 0 =
i
M–2(10)= 0 M = 20 k’
G S c hShear V, found by vertical equilibrium, ΣV=0 (up +)

right of a = left of b V= 0-2 V=-2k
i g h t G Left of a and right of b, shear is zero because there is no beam to resist it (reaction at b
reduces shear to zero). Shear may be checked, considering it starts and stops with zero.
y r Concentrated loads or reactions change shear from left to right of them. Without load
Cop
between a and b (beam DL assumed negligible) shear is constant.
Bending moment M, found by moment equilibrium, ΣM=0 (clockwise +)
at a M = -2(0) = 0 k’
at mid-span M = -2(5) = -10 k’
at b M = -2(10) = -20 k’
The mid-span moment being half the moment at b implies linear distribution. The support
reaction moment is equal and opposite to the beam moment.
Deflection Δ is described later. Diagrams visualize positive and negative bending by
concave and convex curvature, respectively. They are drawn, visualizing a highly flexible
beam, and may be used to verify bending.

Simple beam with uniform load
1 Beam of L= 20 ft span, with uniformly distributed load w = 100 plf
2 Free-body diagram of partial beam x units long
3 Shear diagram
4 Bending moment diagram
To find the distribution of shear and bending along the beam, we investigate the beam at
intervals of 5’, from left to right. This is not normally required.
Reactions R are half the load on each support due to symmetry
R= w L/2 = 100 (20)/2 R= +1000 lbs
Shear force Vx at any distance x from left is found using ∑ V = 0
∑ V = 0; R - w x - Vx = 0; solving for Vx
6
Vx = R - w x
at x = 0’
at x = 5’
V = 0 + Ra = 0 +1000
V = + 1000 - 100 (5)
0 - 2 0 0 V = +1000 lbs
V = +500 lbs
9
at x = 10’ V = + 1000 - 100 (10) V = 0 lbs
at x = 15’
1 9V = + 1000 - 100 (15) V = -500 lbs
erle
at x = 20’ V = + 1000 - 100 (20) V = -1000 lbs
i
Bending moment Mx at any distance x from left is found by ∑ M = 0.
GS c h∑ M = 0; R x – w x (x/2) - Mx = 0;
Mx = R x - w x2/2
solving for Mx
i g h t G at x = 0
at x = 5’
M = 1000 (0) - 100 (0)2/2
M = 1000 (5) - 100 (5)2/2
M = 0 lb-ft
M = 3750 lb-ft
y r at x = 10’ M = 1000 (10)-100 (10)2/2 M = 5000 lb-ft
Cop
at x = 15’ M = 1000 (15)-100 (15)2/2 M = 3750 lb-ft
at x = 20’ M = 1000 (20)-100 (20)2/2 M = 0 lb-ft
Bending is zero at both supports since pins and rollers have no moment resistance.
Since the bending formula Mx= Rx-wx2/2 is quadratic, bending increase is quadratic
(parabolic curve) toward maximum at center, and decreases to zero at the right support.
For simple beams with uniform load the maximum shear force is at the supports and the
maximum bending moment at mid-span (x= L/2) are:
Vmax = R = w L / 2
Mmax = (wL/2)L/2 – (wL/2)L/4 = 2wL2/8 - wL2/8, or
Mmax = wL2 / 8
This formula is only for simple beams with uniform load. Verifying example:
Mmax= wL2/8 = 100(20)2/8 Mmax= +5000 lbs-ft (same as above)

Area Method
The area method for beam design simplifies computation of shear forces and bending
moments and is derived, referring to the following diagrams:
1 Load diagram on beam
2 Beam diagram
3 Shear diagram
4 Bending diagram
The area method may be stated:
• The shear at any point n is equal to the shear at point m plus the area of the load
diagram between m and n.
• The bending moment at any point n is equal to the moment at point m plus the area
6
of the shear diagram between m and n.
2 0 0
The shear force is derived using vertical equilibrium:
0 -
9
∑ V = 0; Vm - w x - Vn = 0; solving for Vn
1 9
erle
Vn = Vm-wx
i
where w x is the load area between m and n (downward load w = negative).
GS c hThe bending moment is derived using moment equilibrium:

∑ M = 0; Mm + Vmx – w x x/2 - Mn = 0;
G
solving for M n
rig h t Mn = Mm+Vmx-wx2/2
C o p y where Vmx – wx2/2 is the shear area between m and n, namely, the rectangle Vm x less
the triangle w x2/2. This relationship may also be stated as
Mn = Mm + Vx, where V is the average shear between m and n.
By the area method moments are usually equal to the area of one or more rectangles
and/or triangles. It is best to first compute and draw the shear diagram and then
compute the moments as the area of the shear diagram.
From the diagrams and derivation we may conclude:
• Positive shear implies increasing bending moment.
• Zero shear (change from + to -) implies peak bending moment
(useful to locate maximum bending moment).
• Negative shear implies decreasing bending moment.
Even though the forgoing is for uniform load, it applies to concentrated load and non-
uniform load as well. The derivation for such loads is similar.

Examples
The following wood beams demonstrate the area method for design and analysis. For
design, a beam is sized for given loads; for analysis, stresses are checked against
allowable limits, or how much load a beam can carry.
Beam design
V Shear diagram.
M Bending diagram.
Δ Deflection diagram.
I Inflection point (change from + to - bending).
Reactions
R= 400plf (24)/2 R = 4800 lbs
Shear
Va = 0
Vbl = 0 -400(5)
- 2 0 0 6 Va = 0 lbs
Vbl = -2000 lbs
0
Vbr = -2000 + 4800 Vbr = +2800 lbs
1 9
Vcl = +2800 - 400(14)
Vcr = -2800 + 4800
9 Vc l= -2800 lbs
Vcr = +2000 lbs
erle
Vd = +2000 - 400(5) Vd = 0 lbs, ok
c h i
Moment
GS
Ma = 0
Mb = 0 -2000 (5)/2 Mb = -5000 lb-ft
i g h t G Mb-c = -5000+ 2800 (7)/2

Mc = +4800 - 2800 (7)/2
Mb-c = +4800 lb-ft
Mc = -5000 lb-ft
r
Md = - 5000+ 2000 (5)/2 Md = 0, ok
Copy Try 4x10 beam

S= (3.5) 9.252/6
Bending stress
S = 50 in3
f b = Mmax/S= 5000 (12)/50 f b = 1200 psi

1200 < 1450, ok
Shear stress
f v = 1.5V/bd=1.5(2800)/[3.5(9.25)] f v = 130 psi
130 > 95, not ok
Try 6x10 beam
f v = 1.5V/bd=1.5 (2800)/[5.5 (9.25)] f v= 83 psi
83 < 95, ok
Note: increased beam width is most effective to reduce shear stress; but increased depth
is most effective to reduce bending stress.

Beam analysis
Reactions
Σ Mc = 0 = +16 Rb - 1000 (20) - 300 (4) 18 - 200 (16) 8
16 Rb = 1000(20) + 300 (4) 18 + 200 (16) 8
Rb = (20000 + 21600 + 25600)/16 Rb = +4200 lbs
Σ Mb = 0 = -16 Rc -1000(4) - 300(4)2 + 200(16)8
16 Rc = -1000(4) - 300(4)2 + 200(16) 8
Rc = (-4000 -2400 + 25600)/16 = Rc = +1200 lbs
Shear
Var = 0 - 1000 Var= -1000 lbs
Vbl = -1000 - 300 (4) Vbl = -2200 lbs
Vbr = -2200 + 4200
Vcl = +2000 - 200(16) = -Rc
- 2 0 0 6 Vbr= +2000 lbs

Vcl= -1200 lbs
0
Vcr = -1200 + 1200 Vcr= 0 lbs
1 9 9
Find x, where shear = 0 and bending = maximum:
erle
Vbr-w2 x = 0; x = Vbr/w2 = 2000/200 x = 10 ft
c h i
Moment
GS
Ma = 0
Mb = 0 + 4 (-1000 -2200)/2 Mb = -6400 lb-ft
G
Mx = -6400 + 10 (2000)/2 Mx = +3600 lb-ft
r i g h t Mc = +3600 + (16-10)(-1200)/2 Mc = 0
y
Section modulus
Cop
S = bd2/6 = (3.5)11.252/6 S = 74 in3
Bending stress
f b = M/S= 6400(12)/74 f b = 1038 psi
1038 < 1450, ok
Shear stress
fv = 1.5V/(bd) = 1.5(2200)/[3.5(11.25)] f v = 84 psi
84 < 95 = ok
Note: stress is figured, using absolute maximum bending and shear, regardless if positive
or negative. Lumber sizes are nominal, yet actual sizes are used for computation.
Actual sizes are ½ in. less for lumber up to 6 in. nominal and ¾ in. less for larger sizes:
4x8 nominal is 3½x7¼ in. actual.
Note: in the above two beams shear stress is more critical (closer to the allowable stress)
than bending stress, because negative cantilever moments partly reduces positive
moments.

Indeterminate beams
Indeterminate beams include beams with fixed-end (moment resistant) supports and
beams of more than two supports, referred to as continuous beams. The design of
statically indeterminate beams cannot be done by static equations alone. However,
bending coefficients, derived by mechanics, may be used for analysis of typical beams.
The bending moment is computed, multiplying the bending coefficients by the total load
W and span L between supports. For continuous beams, the method is limited to beams
of equal spans for all bays. The coefficients here assume all bays are loaded.
Coefficients for alternating live load on some bays and combined dead load plus live load
on others, which may result in greater bending moments, are in Appendix A. Appendix A
also has coefficients for other load conditions, such as various point loads. The equation
for bending moments by bending coefficients is:
6
M=CLW
M = bending moment
C = bending coefficient
0 - 2 0 0
9
L = span between supports
1 9
erle
W = w L (total load per bay)
w = uniform load in plf (pounds / linear foot
S c h
1
i Simple beam
G
2 Fixed-end beam
G
(combined positive plus negative moments equal the simple-beam moment)
ht
3 Two-span beam
r i g 4 Three-span beam
Copy

Flexure Formula
The flexure formula gives the internal bending stress caused by an external moment on a
beam or other bending member of homogeneous material. It is derived here for a
rectangular beam but is valid for any shape.
1 Unloaded beam with hatched square
2 Beam subject to bending with hatched square deformed
3 Stress diagram of deformed beam subject to bending
Referring to the diagram, a beam subject to positive bending assumes a concave
curvature (circular under pure bending). As illustrated by the hatched square, the top
shortens and the bottom elongates, causing compressive stress on the top and tensile
stress on the bottom. Assuming stress varies linearly with strain, stress distribution over
the beam depth is proportional to strain deformation. Thus stress varies linearly over the
2 0 0 6
depth of the beam and is zero at the neutral axis (NA). The bending stress fy at any
distance y from the neutral axis is found, considering similar triangles, namely fy relates
-
0
to y as f relates to c; f is the maximum bending stress at top or bottom and c the distance
9 9
from the Neutral Axis, namely fy / y = f / c. Solving for fy yields
1
erle
fy = y f / c
i
To satisfy equilibrium, the beam requires an internal resisting moment that is equal and
GS c hopposite to the external bending moment. The internal resisting moment is the sum of all
partial forces F rotating around the neutral axis with a lever arm of length y to balance the
external moment. Each partial force F is the product of stress fy and the partial area a on
i g h t G which it acts, F = a fy. Substituting fy = y f / c, defined above, yields F = a y f / c. Since

the internal resisting moment M is the sum of all forces F times their lever arm y to the
y r neutral axis, M = F y = (f/c) Σ y y a = (f/c) Σ y2a, or M = I f/c, where the term Σy2a is
Cop
defined as moment of inertia (I = Σy2a) for convenience. In formal calculus the
summation of area a is replaced by integration of the differential area da, an infinitely
small area:
I = ∫ y2da I = moment of inertia.

The internal resisting moment equation M = I f/c solved for stress f yields
f=Mc/I the flexure formula,
which gives the bending stress f at any distance c from the neutral axis. A simpler form
is used to compute the maximum fiber stress as derived before. Assuming c as
maximum fiber distance from the neutral axis yields:
f=M/S S = I/c = section modulus
Both the moment of inertia I and section modulus S define the strength of a cross-section
regarding its geometric form.

Section modulus
Rectangular beams of homogeneous material, such as wood, are common in practice.
The section modulus for such beams is derived here.
1 Stress block in rectangular beam under positive bending.
2 Large stress block and lever-arm of a joist in typical upright position.
3 Small, inefficient, stress block and lever-arm of a joist laid flat.
Referring to 1, the section modulus for a rectangular beam of homogeneous material
may be derived as follows. The force couple C and T rotates about the neutral axis to
provide the internal resisting moment. C and T act at the center of mass of their
respective triangular stress block at d/3 from the neutral axis. The magnitude of C and T
is the volume of the upper and lower stress block, respectively.
6
C = T = (f/2) (bd/2) = f b d/4.
2 0 0
The internal resisting moment is the sum of C and T times their respective lever arm, d/3,
to the neutral axis. Hence
0 -
1 9 9
M = C d/3 + T d/3. Substituting C = T = f bd/4 yields
erle
M = 2 (f bd/4) d/3 = f bd2/6, or M = f S,
c h i
where S = bd2/6, defined as the section modulus for rectangular beams of homogeneous
GS
material.
G
S = bd2/6
rig h t Solving M = f S for f yields the maximum bending stress as defined before:
C o p y f = M/S
This formula is valid for homogeneous beams of any shape; but the formula S = bd2/6 is
valid for rectangular beams only. For other shapes S can be computed as S = I /c as
defined before for the flexure formula. The moment of inertia I for various common
shapes is given in Appendix A.
Comparing a joist of 2”x12” in upright and flat position as illustrated in 2 and 3 yields an
interesting observation:
S = 2x122/6 = 48 in3 for the upright joist
S= 12x22/6 = 8 in3 for the flat joist.
The upright joist is six times stronger than the flat joist of equal cross-section. This
demonstrates the importance of correct orientation of bending members, such as beams
or moment frames.

Moment of inertia
The formula for the moment of inertia I=∫y2da reveals that the resistance of any
differential area da increases with its distance y from the neutral axis squared, forming a
parabolic distribution. For a beam of rectangular cross-section, the resistance of top and
bottom fibers with distance y = d/2 from the neutral axis is (d/2)2. Thus, the moment of
inertia, as geometric resistance, is the volume of all fibers under a parabolic surface,
which is 1/3 the volume of a cube of equal dimensions, or I= bd (d/2)2/3, or
I = bd3 / 12 (for rectangular beams only)
the moment of inertia of a rectangular beam of homogeneous material. A formal calculus
derivation of this formula is given in Appendix A. The section modulus gives only the
maximum bending stress, but the moment of inertia gives the stress at any distance c
2 0
asymmetrical cross-section, such as T- and L-shapes.
- 0 6
from the neutral axis as f= Mc/I. This is useful, for example, for bending elements of
1
2
9 9 0
Bending stress distribution over beam cross-section
1
Moment of inertia visualized as volume under parabolic surface
erle
3 T-bar with asymmetrical stress: max. stress at c2 from the neutral axis
i
4 Angle bar with asymmetrical stress distribution about x, y, and z-axes: maximum
GS c h resistance about x-axis and minimum resistance about z-axis
i g h t G
y r
Cop

Moment of inertia and shapes
The moment of inertia, a measure of geometric strength and stiffness, is greatly effected
by a beam’s shape. The formula I= ∫y2da reveals that the resisting capacity of fibers
increase is quadratic with vertical distance from the neutral axis. Material far from the
neutral axis increases the resistance capacity; by contrast, material located near the
neutral axis is relatively ineffective. This is visualized here by the capacity of various
beam shapes to resist bending deformation. The moment of inertia is shown here, along
with relative deformations under gravity load. This qualitative, intuitive comparison is
quantified in Beam deflection of this chapter.
1 Efficient upright 2”x12” joist; I= 2(12)3/12 I = 288 in4

2 Inefficient flat 12”x2” joist; I= 12(2)3/12 I = 8 in4
6
3 Efficient wide-flange beam
4 Inefficient cross-shaped beam
0 - 2 0 0
9
Given the same cross-section area, the upright joist has a 36 times greater moment of
1 9
inertia to resist deformation than the flat one. This represents the square of the joist’s
erle
width-to-depth ratio. A similar contrast can be observed between wide-flange and cross-
shaped beams.
c h i
h t G GS
y r i g
Cop

Shear stress
The distribution of shear stress over the cross-section of beams is derived, referring to a
beam part of length x marked on diagrams. Even though horizontal and vertical shear
are equal at any part of a beam, horizontal shear stress is derived here because it is
much more critical in wood due to horizontal fiber direction.
1 Beam, shear and bending diagrams with marked part of length x
2 Beam part with bending stress pushing and pulling to cause shear
3 Beam part with bending stress above an arbitrary shear plane
Let M be the differential bending moment between m and n. M is equal to the shear area
between m and n (area method), thus M = V x. Substituting V x for M in the flexure
formula f= M c /I yields bending stress f= V x c/I in terms of shear. The differential
bending stress between m and n pushes top and bottom fibers in opposite directions,
6
causing shear stress. At any shear plane y1 from the neutral axis of the beam the sum of
- 2 0 0
shear stress above this plane yields a force F that equals average stress fy times the
cross section area A above the shear plane, F = A fy. The average stress fy is found from
0
9
similar triangles; fy relates to y as f relates to c, i.e., fy/y = f/c; thus fy = f y/c. Since f= V x
1 9
c/I, substituting V x c/I for f yields fy = (V x c/I) y/c = V x y/I. Since F = A fy, it follows that
erle
F = A V x y/I. The horizontal shear stress v equals the force F divided by the area of the
i
shear plane;
GS c hV = F/(x b) = A V x y/(I x b) = V A y/(I b)

The term A y is defined as Q, the first static moment of the area above the shear plane
i g h t G times the lever arm from its centroid to the neutral axis of the entire cross-section.
Substituting Q for A y yields the working formula
y r
Cop
v = V Q / (I b) (shear stress)
v = horizontal shear stress.
Q = static moment (area above shear plane times distance from centroid of that area to
the neutral axis of the entire cross-section
I = moment of Inertia of entire cross section
b = width of shear plane
The formula for shear stress can also be stated as shear flow q, measured in force per
unit length (pound per linear inch, kip per linear inch, or similar metric units); hence
q=VQ/I (shear flow)
q = force per unit length

Shear stress in wood and steel beams
Based on the forgoing general derivation of shear stress, the formulas for shear stress in
rectangular wood beams and flanged steel beams is derived here. The maximum stress
in those beams is customarily defined as fv instead of v in the general shear formula.
1 Shear at neutral axis of rectangular beam (maximum stress),
Q = Ay = (bd/2) d/4, or
Q = bd2/8 (Note: d2 implies parabolic distribution)
I = bd /12, hence
3
v = V Q / I b = V (bd2/8) / [(bd3/12)b] = fv, or

fv = 1.5 V / (bd)
- 2 0 6
Note: this is the same formula derived for maximum shear stress before
0
Shear stress at the bottom of rectangular beam. Note that y= 0 since the centroid
1 9 9 0
of the area above the shear plane (bottom) coincides with the neutral axis of the
entire section. Thus Q= Ay = (bd/2) 0 = 0, hence
erle
v = V 0/(I b) = 0 = fv, thus
i
fv = 0
GS c hNote: this confirms an intuitive interpretation that suggests zero stress since no fibers
below the beam could resist shear
i g h t G 3 Shear stress at top of rectangular beam. Note A = 0b = 0 since the depth of the
shear area above the top of the beam is zero. Thus
r
Q = Ay = 0 d/2 = 0, hence v = V 0/(I b) = 0 = fv, thus
Copy fv = 0
Note: this, too, confirms an intuitive interpretation that suggests zero stress since no
fibers above the beam top could resist shear.
4 Shear stress distribution over a rectangular section is parabolic as implied by the
formula Q=bd2/8 derived above.
5 Shear stress in a steel beam is minimal in the flanges and parabolic over the web.
The formula v = VQ/(I b) results in a small stress in the flanges since the width b of
flanges is much greater than the web thickness. However, for convenience, shear
stress in steel beams is computed as “average” by the simplified formula:
fv = V / Av
fv = shear stress in steel beam
V = shear force at section investigated
Av = shear area, defined as web thickness times beam depth

Shear stress in wood I-beam
Since this is not a rectangular beam, shear stress must be computed by the general
shear formula. The maximum shear stress at the neutral axis as well as shear stress at
the intersection between flange and web (shear plane As) will be computed. The latter
gives the shear stress in the glued connection. To compare shear- and bending stress
the latter is also computed.
1 Beam of L= 10 ft length, with uniform load w= 280plf (W = 2800 lbs)
2 Cross-section of wood I-beam
Shear force V= W/2 = 2800/2 V = 1400 lbs
Bending moment M=WL/8 = 2800(10)/8 M = 3500 lb’
For the formula v= VQ/(Ib) we must find the moment of inertia of the entire cross-section.
6
We could use the parallel axis theorem of Appendix A. However, due to symmetry, a
- 2 0 0
simplified formula is possible, finding the moment of inertia for the overall dimensions as
rectangular beam minus that for two rectangles on both sides of the web.
0
9
I = (BD3– bd3)/12 = [6(10)3– 2(2)63]/12
1 9 I = 428 in4
rle
Bending stress fb= Mc/I = 3500(12)5/428 fb = 491 psi
i e
491 < 1450, ok
G S c hNote c= 10/2 = 5 (half the beam depth due to symmetry)

Static moment Q of flange about the neutral axis:
i g h t G Q = Ay = 6(2)4
Shear stress at flange/web intersection:
Q = 48 in3
y r v = VQ/(Ib) = 1400(48)/[428(2)] v = 79 psi
Cop
Static moment Q of flange plus upper half of web about the neutral axis
Q = ΣAy = 6(2)4 + 2(3)1.5 Q = 57 in3
Maximum shear stress at neutral axis:
v = VQ/(Ib) = 1400(57)/[428(2)] v = 93 psi <95, ok
Note: Maximum shear stress reaches almost the allowable stress limit, but bending
stress is well below allowable bending stress because the beam is very short. We can
try at what span the beam approaches allowable stress, assuming L= 30 ft, using the
same total load W = 2800 lbs to keep shear stress constant:
M= WL/8 = 2800(30)/8 M = 10500 lb-ft
fb = Mc/I = 10500(12)5/428 fb = 1472 psi
1472 >1450, not ok
At 30 ft span bending stress is just over the allowable stress of 1450 psi. This shows that
in short beams shear governs, but in long beams bending or deflection governs.

Shear stress in steel beam
This beam, supporting a column point load of 96 k over a door, is a composite beam
consisting of a wide-flange base beam with 8x½ in plates welded to top and bottom
flanges. The beam is analyzed with and without plates. As shown before, for steel
beams shear stress is assumed to be resisted by the web only, computed as fv = V/Av.
The base beam is a W10x49 [10 in (254 mm) nominal depth, 49 lbs/ft (6.77 kg/m) DL]
with a moment of inertia Ixx= 272 in4 (11322 cm4) (see Appendix). Shear in the welds
connecting the plates to the beam is found using the shear flow formula q = VQ/(I).
1 Beam of L= 6 ft (1.83 m) span with P = 96 k point load
2 Composite wide-flange beam W10x49 with 8x½ inch stiffener plates
Shear force V = P/2 = 96/2 V = 48 k
Bending moment M = 48(3) M = 144 k’
6
Wide-flange beam
Shear area of web Av = web thickness x beam depth
Av = 0.34(10)
0-2 0 0 Av = 3.4 in2
1 9 9
Shear stress fv = V/Av = 48/3.4 fv = 14 ksi
14 < 14.5, ok
erle
Bending stress fb = Mc/I = 144(12)5.5/272 fb = 35 ksi
c h i 35 >22, not ok
GS
Since the beam would fail in bending, a composite beam is used.
G
Composite beam
rig h t Moment of inertia I=Σ(Ioo+Ay2) (see parallel axis theorem in Appendix A)

I= 272+2(8)0.53/12+2(8)0.5(5.25)2 I = 493 in4
C o p y Bending stress fb= Mc/I = 144(12)5/493
Since the shear force remains unchanged, the web shear stress is still ok.
fb = 19 ksi
19 < 22, ok
Shear flow q in welded plate connection

Q= Ay = 8(.5)5.25 = 21 in3
qtot = VQ/I = 48(21)/493 qtot = 2 k/in
Since there are two welds, each resists half the total shear flow
q = qtot/2 q = 1k/in
Assume ¼ in weld of 3.2 k/in * strength 1 < 3.2, ok
* see AISC weld strength table
Note: in this steel beam, bending is stress is more critical than shear stress; this is typical
for steel beams, except very short ones.

Deflection
To satisfy stiffness, beam deflection must be limited by code or other factors. For
example, to prevent cracks in plaster, codes require deflection to be not more than L/360
for LL or L/240 for combined LL+DL. Excessive deflection may also be unsightly or
cause damage to non-load-bearing partitions. Therefore, beams may be oversized for
strength to limit deflection.
Beam deflection is caused by both bending and shear, yet, except for very short beams,
shear deflection is typically very small and may be ignored.
1 Simple beam under uniform load
2 Bending deflection of simple beam under uniform load
3 Shear deflection of simple beam under uniform load
4 Bending deflection of cantilever beam under point load
6
5 Shear deflection of cantilever beam under point load
on the following pages as:
0 - 2 0 0
Referring to 4, elastic bending deflection of a cantilever beam under point load is derived
9
Δ = PL3 / (3EI)
1 9
erle
E = modulus of elasticity, I = moment of inertia
c h i
Referring to 5, shear deflection is defined by the formula:
GS
Δ = 6PL/(5AG)
i g h t G A = cross-section area, G = shear modulus

Shear deflection is not derived, since it is negligible and ignored for most beams. The
y r above formulas show bending deflection increases with the third power of L, but shear
Cop
deflection increases linearly with L. Shear deflection is equal to shear stress (V/A = P/A)
divided by the shear modulus G, modified by 6/5 since shear stress is non-linear over the
beam depth.
Referring to 4 and 5, the following highlights the correlation of beam length with shear-
and bending deflection. Assuming a 4x6 in (102x152 mm) cantilever steel beam with
P=8 k (36 kN), E=30,000 ksi (206,850 MPa), G=12,000 ksi (82,740 MPa), I=4x63/12=72
in4 (29969x10-3 mm4), A = 24 in2 (16 cm2)
If the beam length is L=60in (152 cm), bending deflection is Δ=(8)603/[3(30000)72]=.27in
(7mm), but shear deflection is only Δ=6(8)60/[5(12000)24]=.002 in (.05 mm). Thus shear
deflection is less than one percent of bending deflection. However, if the beam length
equals the beam depth, L=6 in (152 mm), then the bending deflection is reduced to
Δ=.00027 in (.007 mm) and shear deflection to Δ=.0002 in (.005 mm); which is about
equal to the bending deflection. This confirms, shear deflection is insignificant and may
be ignored for beams of typical length, but approaches bending deflection when the
beam length is reduced to the beam depth.

Moment-area method
The moment-area method for deflection was developed in 1873 by Charles Green of the
USA independent of a similar method developed in 1868 by Otto Mohr of Germany. The
derivation of the moment-area theorem is based on fig. 1 and 2, showing part of a
deformed beam and its elastic curve AB, respectively; assuming small deformations and
constant elastic modulus E and moment of inertia I. Referring to fig. 1, let GB be parallel
to FO, then FG=AB, the unstressed length, and GH/AB=ε, the unit strain. Since the
elastic modulus is E=f/ε, f=Eε, or f=E GH/AB; but GH/AB=c/r, due to similar triangles.
Substituting c/r for GH/AB yields f=Ec/r and, since f=Mc/I (where M= bending moment -
see flexure formula), Ec/r=Mc/I, hence E/r=M/I, or 1/r=M(EI).
Referring to fig. 2, with angles dφ and θ measured in radians, dφ is the angle of the radii
6
at m and n and between the tangents to those radii. The length dx=r dφ and dφ/dx=1/r =
θ=∑dφ = ∑Mdx/(EI), or
0 - 2 0 0
M(EI) (as derived above), or dφ=M dx /(EI). The sum of dφ between A to B is
θ= Am/(EI)
1 9 9
erle
Am = ∑Mdx, the area of the bending moment diagram between A and B. Hence, the
i
theorem for the beam slope may be stated as follows:
GS c h The angle θ between the tangents of points A and B on the elastic curve of a
beam is the moment diagram area between A and B, divided by EI.
i g h t G This theorem can be used to find the elastic slope at any point of a beam. The theorem
for deflection (usually of greater interest) is derived next.
y r
Cop
The angle between the tangents at m and n on the elastic curve is dφ and the vertical
displacement between these tangents at A is xdφ. Therefore, the displacement between
A and the tangent at B is Δ=∑xdφ=∑xMdx/(EI), or
Δ= x Am/(EI)
where Am = ∑Mdx, area of the bending moment diagram between A and B and times the
lever arm from A to the centroid of the bending moment diagram between A and B.
Hence, the deflection theorem may be stated as follows:
The vertical displacement Δ of the tangent at B on the elastic curve equals the
moment of the area of the bending diagram between A and B times the lever-arm
x from its centroid to A, divided by EI.
This theorem can be applied to compute beam deflection as shown on the following
pages. The above derivation considers only bending deformation, and ignores shear
deformation, which is insignificant as shown before, and can be ignored for most beams.

Deflection formulas
Based on the moment-area method, the following formulas for slope and deflection are
derived for beams of common load and support conditions. Additional formulas are
provided in Appendix A. Although downward deflection would theoretically be negative, it
is customary to ignore the sign convention and define up- or downward deflection by
inspection. The angle θ is the slope of the tangent to the elastic curve at the free end for
cantilever beams and at supports for simple beams; Δ is the maximum deflection for all
cases. As derived before, θ is the area of the bending moment diagram divided by EI,
the elastic modulus and moment of inertia, respectively; Δ equals θ times the lever-arm
from the centroid of the bending moment diagram between zero and maximum deflection
to the point where θ is maximum.
θ = (PL)(L/2)/(EI), Δ= θ 2/3 L
6
1 Cantilever beam with point load;
θ= 1/2 PL2/(EI)
0 - 2 0 0
9
Δ= 1/3 PL3/(EI)
2
rle 1 9
Cantilever beam with uniform load; θ = (WL/2)(L/3)/(EI), Δ= θ 3/4 L
e
θ= 1/6
i
WL2/(EI)
G Sc h3
Δ= 1/8 WL3/(EI)
Simple beam with point load; θ = (PL/4)(L/4)/(EI), Δ= θ 1/3 L
i ght G θ= 1/16 PL2/(EI)
py r Δ= 1/48 PL3/(EI)
Co
4 Simple beam with uniform load; θ = (WL/8)(2/3 L/2)/(EI), Δ= θ 5/16 L
θ= 1/24 WL2/(EI)
Δ= 5/384 WL3/(EI)

Steel beam with point loads
This steel beam supports joists that span 32 feet between beams and carry a roof load of
50 psf (30 psf DL and 20 psf LL). The joist reactions generate point loads of P=50
psf(10ft)16ft/1000), P= 8 k beam load. The beam is designed for stress, then verified for
deflection and redesigned if necessary.
1 Beam diagram.
2 Load diagram abstraction
Note:
Load at the beam supports is ignored since it is directly supported by columns and hence
has no effect on shear, bending moment, or deflection
V Shear diagram.
M Bending moment diagram.
Δ
Shear:
Deflection diagram.
- 2 0 0 6
Va = Vbl = R = 2(8)/2
1 9
Vbr = Vcl = 8-8
9 0 Va = Vbl = 8 k
Vbr = Vcl = 0 k
erle
Vbr = Vc = R = 0-8 Vdr = Vc =-8 k
i
Bending moment:
G S c h
Mmax= 8 (10)
Section modulus S and moment of inertia I (from Appendix D):
Mmax = 80 k’
i g h t G S= M/Fb= 80 k’(12”)/22 ksi

Try W10x45, S= 49.1 in3
S = 44 in3
49.1 > 44, ok
r
I =248 in4
Co p y Deflection (see Appendix A for formula):

L= 30’ (12”) = 360 in
Δmax = (23/684) PL3/(EI) = [(23/684)8(360)3] / [(30000)248]
Δmax = 1.7 in
Δall=360/240 = 1.5 in 1.5 < 1.7, not ok
Try W18x35, S= 57.6 in3 57.6 > 44, ok

I = 510 in4
Deflection:
Δmax= (23/684) 8 (360)3 / [(30000)510] Δmax = 0.8 in
Δall=2 in 1.5 > 0.8, ok
Note: the W10x45 deflects too much, and, with a span/depth ratio of 36:1, is too shallow;
but W18x35 has the recommended 20:1 ratio, is lighter and, hence, more economical.

Steel beam with mixed load
This steel beam, too, supports joists that span 32 feet between beams and carry a roof
load of 50 psf (30 psf DL and 20 psf LL). The joist reactions generate point loads of
P=50 psf(10ft)16ft/1000), P = 8 k beam load. In addition, the beam carries a uniform
dead load of 0.4 klf (the beam’s own weight plus fire proofing and cladding). The beam is
designed for stress, then verified for deflection and redesigned if necessary.
1 Beam diagram
2 Load diagram abstraction
Note: load at the beam supports is ignored since it has no effect on shear, bending
moment, or deflection
V Shear diagram
M Bending moment diagram
Δ
Shear:
Deflection diagram
- 2 0 0 6
Va = R= [2(8)+0.4(30)]/2
1 9
Vbl = 14-0.4(10)
9 0 Va = 14 k
Vbl = 10 k
erle
Vbr = 10-8 Vbr = 2 k
Vcl = 2-0.4(10) Vcl= -2 k
c h i
Vcr = -2-8 Vcr= -10 k
GS
Vd = -10-0.4(10) Vdr= -14 k
Bending moment:
i g h t G Mmax= 10(14+10)/2 + 5(2)/2 Mmax = 125 k’
r
Section modulus S and moment of inertia I (from Appendix D):
Copy S= M/Fb= 125 k’(12”)/22 ksi

Try W18x40, S= 68.4 in3
Deflection (see Appendix A for formula):

S = 68 in3
68.4 > 68, ok
I = 612 in4
L= 30’ (12”) = 360 in

W= wL = 0.4 (30) = 12 k
Δmax = (23/684) PL3/(EI) + (5/384) WL3/(EI) *
Δmax = [(23/684)8(360)3 + (5/384) 12(360) 3] / [(30000)612]
Δmax = 1.1 in
Δall=360/240 = 1.5 in 1.5 > 1.1, ok
* Superimposition of equations for point load and uniformly distributed load

Typical beam diagrams
Deflection, shear, and bending diagrams are shown here for typical beams. The beam
with deflection and load diagrams are drawn on top with shear and bending diagrams
shown below. With experience, these diagrams may be drawn by visual inspection prior
to computing. This is useful to verify computations and develop an intuitive sense and
visualization regarding shear and bending on beams. The deflection diagram is drawn,
visualizing the deflection of a thin board, flexible ruler, or similar device. It is drawn
grossly exaggerated to be visible. The shear diagram is drawn at a convenient force
scale left to right, starting with zero shear to the left of the beam. Downward uniform load
yields downward sloping shear. Downward point loads are drawn as downward offset,
and upward reactions yield upward offset. Bending diagrams are drawn, considering the
area method; namely, bending at any point is equal to the area of the shear diagram up
6
to that point. Both, shear and bending must be zero to the right of the right beam end. To
- 2 0 0
satisfy this, requires a certain amount of forward thinking and, in complex cases, even
working backward from right to left as well as left to right.
0
1
2
9
Cantilever beam with point load
1
Cantilever beam with uniform load
9
erle
3 Cantilever beam with mixed load
i
4 Simple beam with point loads
GS c h5
6
7
Simple beam with uniform load
Simple beam with mixed load
Beam with one overhang and point load
i g h t G 8
9
Beam with one overhang and uniform load
Beam with one overhang and mixed load
y r 10 Beam with two overhangs and point loads
Cop
11 Beam with two overhangs and uniform load
12 Beam with two overhangs and mixed load

ASD (Allowable Stress Design) and LRFD (Load and Resistance Factor Design) are two
8 design and analysis methods currently used for structural design. ASD is the classic
method used since the inception of structural design and sometimes referred to as
working stress design. LRFD is a new method, increasingly promoted by codes. The
difference of the two methods is essentially in the way they consider the issue of safety:
ASD uses actual loads to design members for allowable stress that is reduced from
ultimate strength or yield stress by a safety factor. By contrast LRFD assigns safety to
the load, increasing actual service load by a load factor to design members for stress that
is close to the ultimate strength. The load factors provide a more rational safety because
ASD, LRFD, M a sonry, dead load is more predictable than live load and therefore has a smaller load factor
LRFD is similar to the Strength Method or Ultimate Strength Method that has been used
a nd Conc re t e De sign
6
for concrete design since about 1960. The two methods are briefly introduced below and
masonry design ASD is still more common.
0 - 2 0 0
demonstrated for masonry design (ASD) and concrete design (LRFD). At present, for
1 9 9
c h i erle
h t G GS
y r i g
Cop
8-1 DESIGN METHODS ASD, LRFD, Masonry, and Concrete Design

Wood
ASD (Allowable Stress Design) Base values for Douglas Fir-Larch 2”x5” (5x13 cm) or greater for allowable stress:
Allowable stress design, also known as working stress design, was the traditional method bending (Fb), tension (Ft), compression (Fc), compression normal to grain (FC⊥),
in general use before the advent of the LRFD method. The ASD method is based on horizontal shear (Fv), and elastic modulus (E).
service loads as defined by codes. Structural members are designed to resist such loads
without exceeding allowable stresses, allowable deflections, and lateral drift. Allowable Grade Fb Ft Fc Fc⊥ Fv E units
stresses are based on ultimate strength or yield stress, reduced by safety factors. The Select 1,500 1,000 1,100 625 85 1,600,000 psi
safety factors depend on the consistency of a given material and the type of stress. For structural: 10.3 6.9 7.6 4.3 0.6 1,1032 MPa
example, allowable axial tensile stress for steel is 60 % of the yield stress (Fa= 0.6Fy). No. 1: 1,200 625 1,000 625 85 1,600,000 psi
Allowable deflections for horizontal span members shall not exceed Δ=L/240 for 8.2 4.3 6.9 4.3 0.6 1,1032 MPa
combined dead and live load and No. 2: 700 475 1300 625 85 1,300,000 psi
4.8 3.3 9.0 4.3 0.6 8,964 MPa
Δ=L/360 for live load only
Steel
Δ = maximum deflection
The table gives yield stress (Fy), ultimate strength (Fu), allowable stress for bending (Fb),
6
L = span
0
compression (Fc), tension (Ft), and shear (Fv), elastic modulus (E)
The lateral drift of vertical structures shall not exceed a fraction of the height (Maximum
drift is typically 0.5 % of height).
Steel grade
ASTM A36
Fy
36
Fu
58-80
Fb
22
9 0 - 2 0
Fc Ft
22
Fv
14.5 29,000
E ksi
ksi
9
248 400-550 150 150 100 200,000 MPa
ASD Load combinations
ASTM A572
erle 1 50
345
65
450
30
210
30
210
20
140
29,000
200,000
ksi
MPa
i
Based on the 1997 UBC structures and all portions thereof shall resist the most critical
h
Masonry
c
effects resulting from the following combinations of loads:
GS
Allowable compressive stress Fa, for masonry with special inspection is 25% of
D specified strength ƒ’m by the ASD method; reduced for slenderness. Specified
G
Compressive strength ƒ’m is based on compressive strength of masonry units and
t
D +L + (Lr or S)
h
mortars type M, S, N.
g
D + (W or E/1.4)
0.9D ± E/1.4
y r i Type Concrete masonry (ksi) Clay brick masonry (ksi)
Cop
D + 0.75[L + (Lr or S) + (W or E/1.4)] Unit strength 1.9 2.8 3.75 4.8 4 6 8 10 12 14
ƒ’m (M or S) 1.5 2 2.5 3 2 2.7 3.35 4 4.7 5.3
D = Dead load ƒ’m (N) 1.35 1.85 2.35 2.8 1.6 2.2 2.7 3.3 3.8 4.4
E = Earthquake load
Type Concrete masonry (MPa) Clay brick masonry (MPa)
L = Live load
Unit strength 13 19 26 33 28 41 55 69 83 97
Lr = Roof live load
ƒ’m (M or S) 10 14 17 21 14 19 23 28 32 37
S = Snow load
ƒ’m (N) 9 13 16 19 11 15 19 23 26 30
W = Wind load
Concrete
Most of this book is based on ASD By working stress method, allowable stresses are based on compressive strength ƒ’c.
Typical compressive strengths range from 2 to 6 ksi (14 to 41 MPa)
Allowable stress is defined by a material’s ultimate strength or yield strength and a
factor of safety. Building codes and trade associations provide allowable stress for Allowable compressive stress 0.40 ƒ‘c
various materials and grades of materials, which may also depend on duration of load. Allowable compressive bending stress 0.45 ƒ‘c
Allowable wood stress also depends on temperature, moisture content, size, and if a Allowable shear stress without reinforcing: beam 1.1 ƒ‘c1/2
member is single or repetitive, like closely spaced joists. Relevant factors regarding joist 1.2 ƒ‘c1/2
allowable stress are briefly introduced here and further described later in this chapter. footing & slab on grade 2.0 ƒ‘c1/2

Since factored loads are based on statistical probability and extensive tests, the LRFD
LRFD (Load and Resistance Factor Design) method usually results in smaller members than the ASD method, but the LRFD method
LRFD is a new method increasingly promoted by building codes. It is similar to the is more complex and requires stiffness (deflection and drift) to be computed by actual
strength method used for concrete design since the 1960th. LRFD is based on factored service loads rather than factored loads. Also, since the LRFD method considers
loads (amplified service loads) and nominal resistance (reduced ultimate strength). strength rather than stress, the results cannot be verified for allowable stress as in ASD.
Safety factors are assumed by factored load, rather than allowable stress as in ASD. To verify results requires a second analysis by ASD. This represents a challenge for
Typical factored loads are 1.2 dead load and 1.6 live load. The LRFD design method is future refinement of the LRFD method.
essentially defined by the equation:
φ Design Strength ≥ Required Resistance Example: roof rafters
φ = Resistance factor (φ <1) Roof steel rafters, sloping 2:12, spaced 10’, are subject to the following loads:
The resistance factor depends on the material and type of stress, based on reliability and Dead load D = 40 psf
consistency of tests (low reliability = low φ) Snow load S = 30 psf
Resistance factors φ Wind load W = 15 psf (downward)
6
Material Steel
Stress types
Bending
Concrete *
0.9
Masonry *
0.8
Wood
0.85
Limit states
Yielding
φ
0.9
Find the maximum load effect per liner foot
0 - 2 0 0
9
Shear 0.85 0.6 0.75 D = 40 psf (10’) D = 400 plf
Tension Rupture 0.75
1 9
L = 30 psf (10’) L = 300 plf
erle
Compression 0.75 spiral 0.65 0.9 Compression 0.85 w = 15 psf (10’) w = 150 plf
i
0.70 tied and buckling
h
1.4D
c
Stability 0.85
GS
1.4(400 plf) 560 plf
* Strength design (similar to LRFD)
1.2D + 1.6L + 0.5(Lr or S)
G
LRFD load combinations 1.2(400 plf) + 0.5(300 plf) 630 plf
rig h t
Based on the 1997 UBC structures and all portions thereof shall resist the most critical
effects resulting from the following combinations of factored loads:
1.2D + 1.6(Lr or S) + (f1 L or 0.8W)
1.2(400) + 1.6(300 plf) + 0.8(150 plf) 1080 plf
1.4D
C o p y
1.2D + 1.6L + 0.5(Lr or S)
1.2D + 1.6(Lr or S) + (f1 L or 0.8W)
1.2D + 1.3W + f1 L + 0.5(Lr or S)
1.2D + 1.3W + f1 L + 0.5(Lr or S)
1.2(4000 plf) + 1.3(150 plf) + 0.5(300 plf)
1.2D + 1.0E + (f1 L + f2 S)
1.2 (400 plf) + ( 0.7(300 plf)
1905 plf
690 plf
0.9D ± (1.0E or 1.3W)
1.2D + 1.0E + (f1 L + f2 S)
0.9(400 plf) +1.3(150 plf) 555 plf
0.9D ± (1.0E or 1.3W)
D = Dead load Governing load effect wu = 1905 plf
E = Earthquake load
L = Live load
Lr = Roof live load
S = Snow load
W = Wind load
f1 = 1.0 for floors of public assembly, live load >100 psf and garage live load
0.5 for all other live loads
f2 = 0.7 for roofs that don’t shed snow
0.2 for all other roofs

Masonry Design (ASD)
Allowable masonry stresses require special inspection as defined by bulding codes.
Allowable stresses are one half without special inspection
1 Beam of homogeneous material resistes gravity bending with maximum top
compression, maximum bottom tension and zero stress at the neutral axis
2 Masonry beam resists only compression and steel rebars resist tension. The
stiffness difference of masonry and steel are adjusted by the elastic ratio n = Es / Em
Es = Elastic modulus, steel Es = 29,000 ksi
Em. = Elastic modulus. masonry Em.= 750 f’m
n = Elastic ratio (steel / masonry) n = Es / Em
f’m = Specified masonry compressive strength f’m = 1.5 to 5 ksi
Fb = Allowable masonry bending stress Fb=f’m/3, max.2 ksi *
Fs = Allowable rebar stress: Fs=0.5Fy, max. 24 ksi *
Fs = Allowable stirrup stress:
2 0
Fv = Allowable shear stress if masonry resist all shear
- 0 6 Fs=0.4Fy, max. 24 ksi *

Fv =(f’m)1/2, max.50 psi *
0
Fv = Allowable shear stress if steel resist all shear Fv =3(f’m)1/2, max.150 psi *
*
9 9
1
(Sample tests at start of construction and for every 5000 ft2 of masonry)
erle
b = Beam width
i
d = Effective depth (top of beam to centroid of reinforcing steel)
GS c hh = beam depth (top face to bottom face)

kd = Depth of triangular compression stress block
jd = Moment arm, d-kd/3 (distance from tension to compression centroids)
i g h t G fb = Maximum compressive bending stress

fs = Tensile stress in reinforcing steel
y r As = Cross section area of reinforcing steel
Cop
p = Ratio of steel area / beam cross section, p = As/bd (0.02% to 2.88%)
Referring to diagram 2 the following equations are derived:
For Balanced beams (with enough reinforcing so that steel and masonry reach their
respective limits simultaneously), kd is defined by similar triangles:
kd / d = fb / ( fb+ fs /n)
k = 1 / [1+ fs / (n fb)]
Based on the k-factor other factors are deirved:
Resiting lever arm jd = d - kd/3
j-factor j = 1 - k/3
Resisting moment M = ½ fb b kd jd = ½ fb kj bd2
M = R bd2
Resistance factor R = ½ fb kj
Max. masonry stress fb = 2M / (kj bd2)
Required steel area As = M / (Fs jd)
Steel stress fs = M / (As jd)

Example: masonry beam design
Design a simply supported brick masonry beam
Assume:
L =16’, b = 10”, specified compressive strenght f’m = 1500 psi, with special inspection,
Fb = 1500/3 = 500 psi, grade 60 steel, Fs = 24 ksi
Dead load estimate DL = 300 plf
Live koad estimate LL = 500 plf
w = 800 plf
Bending moment M = w L2/8 = 800(16)2/8 M = 25,600 #’
Elastic modulus
Em = 750 f’m = 750(1500)/1000 Em = 1,125 ksi
Elastic ratio
n = Es/Em = 29,000 ksi / 1875 ksi n = 25.8
6
k = 1/[1+ fs / (n fb)] = 1/[1+24,000/(25.8 x 500]) k = 0.35
j = 1 – k/3 = 1 – 0.35/3
Resistance factor
0 - 2 0 0 j = 0.88
R = ½ fb kj = ½ 500 x0.35 x 0.88
1 9
Effective depth required, M = R bd2)
9 R = 77
erle
d = M/bR = 25,600x12"/(10x77) d = 20”
c h i Beam depth h = d+4” = 20+4 (4” for rebar + cover (adjist for modules) h = 24”
GS
Required steel area
i g h t G As = M / (Fs jd) = 25,600(12”) / (24,000 x 0.88 x 20)

Use 1 # 8 bar, As = π (0.5) 2
As = 0.72 in2
As = 0.79 in2, OK
y r Example: masonry beam analysis
Cop
Assume:
Simple beam, L=10’, b=8”, d=32”, specified compressive strenght f’m = 1500 psi, without
special inspection, Fb =1/2x1500/3 = 250 psi, grade 40 steel, Fs = 20 ksi. 1 # 6 rebar.
Rebar diameters Cross-section areas Dead load estimate DL = 600 plf
Size in in mm in2 mm2 Live koad estimate LL = 900 plf
#3 3/8 0.375 9.5 0.11 71 w = 1500 plf
#4 4/8 0.500 12.7 0.20 129 Bending moment M = w L2/8 = 1500(10)2/8 M = 18,750 #’
#5 5/8 0.625 15.9 0.31 200 k-factor
#6 6/8 0.750 19.1 0.44 284 k = 1/[1+ fs / (n fb)] = 1/[1+20,000/(25.8 x 250]) k = 0.24
#7 7/8 0.875 22.2 0.60 387 j = 1 – k/3 = 1 – 0.24/3 j = 0.92
#8 8/8 1.000 25.4 0.79 510 Max. masonry stress
#9 1.128 28.7 1.00 645 fb = 2M / (kj bd2) = 2x187503x12” / (0.24x0.92x8x322) fb = 248 psi < 250, OK
#10 1.270 32.3 1.27 819
#11 1.410 35.81 1.56 1006 Steel cross section ares
#14 1.693 43.00 2.25 1452 Steel stress
#18 2.257 57.33 4.00 2581 fs = M / (As jd) =18750 x 12” / (0.44x0.92x32) fs = 17,370 psi < 20,000, OK

Shear reinforcing
Bending members are subject to shear that requires reinforcing to prevent diagainal
cracks caused by the tensile components of shear at 45 degrees. Vertical sterrups
provide shear reinforcing. Maximum stirrup spacing of d/2 prevents shear cracks. Thus:
V = Vm+AvFs d/s
V = maximum shear
Vm = shear resisted by masonry
Av = Cross section area of shear reinforcing
Fs = Allowable steel stress
d = effective beam depth
s = stirrup sacing
Empirical UBC formuals
UBC assumes 2 conditions: 1) all shear reisted by masonry; 2) all shear resisted by steel
6
Allowable shear stress if masonry resists all shear Fv = f , max.50 psi *
'
0
m
Allowable shear stress if steel resists all shear
0-2 0 , max.150 psi *

'
Fv = 3 fm
9
* Allowable stresses are one half without special inspection
1 9
Shear resisted by stirrups (Vm ignored) V = Av Fs d / s
erle
Shear resisted by masonry V = Fv bjd
Fv = Allowable masonry shear stress
c h i
Computed shear stress (estimate j = 0.9) fv = V / bjd
GS
Stirrup spacing s = Av Fs / bfv, max. s = d/2
1 Masonry beam with uniform load
G
2 Shear diagram
r i g h t 3 Beam cross section

A Linear bars resist ensile bending stress
Copy b Beam width

C Stirrups (resist shear stress)
d Effective depth (top of beam to steel centroid
E Zone requiring shear reinforcing
h depth of beam
Example: masonry beam
Assume: simple beam, L = 10’, b = 8”, d = 32”, specified strenght f’m = 1500 psi, without
special inspection, Fv =½(1500)1/2 = 19 psi, grade 40 steel, Fs = 16 ksi
DL+LL w = 1500 plf
Max shear V = wL/2 = 1500 x10/2 V =7500 #
Shear stress fv = V / bjd = 7500/ 8x0.9x 32 fv = 33 psi
Stirrup spacing s = Av Fs / bfv= 0.2x16/8x33 s = 12”
Note: for CMU, s would need to be ajusted a multiple of 8” modules
Shear resisted by masonry Vm=Fv bjd =19x8x0.9x32 Vm = 4378 #
Shear resisted by steel Vs = V-Vm = 7500-4378 Vs = 3122 #
Zone requiring shear reinforcing E = (L/2)Vs/V = 5’x3122/7500 E = 2.1’

Shear walls
Shear wall reinforcement to resist lateral load is required in seismic zones 2 to 4. The
reinforcement bars must be provided in both in both vertical and horizontal directions.
Horizontal bars must be continuous or spliced at intersections and wall corners.
Minimum reinforcement is required as follows.
Seismic areas require horizontal and vertical rebars of at least 0.2% of the wall cross
section area. Bars in either direction may be 0.1% but shall be at least 0.07% with the
remaining 0.13% in perpendicular direction. The greater percentage of bars should run
in direction of primary span, normally vertical from floor to floor or roof. Bars shall be
arranged as shown in 1: vertical and horizontal bars spaced maximum 4 ft (1.2 m); bars
around all openings, the top bar extending at least 24 in (60 cm) or 40 bar diameters
beyond openings; on top and bottom of walls; and at structurally connected floors and
6
roofs. Graph 2 gives bar spacing for 0.1 % reinforcing of various wall sizes.
- 2 0 0
Moderate seismic areas requires rebars with cross-section of min. 0.2 in2 (129 mm2), #
0
9
4 bars, arranged as follows: vertical bars at 4 ft (1.2 m); horizontal bars spaced 10 ft (3
1 9
m); bars around all openings extending at least 24 in (60 cm) or 40 bar diameters beyond
erle
openings; bars on top and bottom of walls; and at structurally connected floors and roofs.
i
1 Wall elevation with reinforcing bar layout for seismic zones 2 to 4
GS c h2 Bar size and spacing for 0.1% reinforcing of wall cross-section area
A Vertical bars, spaced maximum 4 ft or 6 times the wall thickness
i g h t G B
C
Horizontal bars spaced max. 4 ft in high seismic areas; 10 ft in moderate areas
Bars at openings, extending min. 2’ or 40 bar diameters beyond opening
y r D Horizontal bars at top and base of wall
Cop
E Bars at structurally connected floors and roof
s Spacing of reinforcing bars, sizes #3 to #7 (max. 6 times bar diameter)
t Wall thickness
Rebar diameters Cross-section areas

Size in in mm in2 mm2
#3 3/8 0.375 9.5 0.11 71
#4 4/8 0.500 12.7 0.20 129
#5 5/8 0.625 15.9 0.31 200
#6 6/8 0.750 19.1 0.44 284
#7 7/8 0.875 22.2 0.60 387
Example: 8” CMU wall
# 4 bar spacing (from graph) s = 24”
# 5 bar spacing, s = 38”, adjust to multiple 8” ≤ 38” s = 32”
# 6 bar spacing, s = 48” max. s = 48”

Allowable shear stress for reinforced masonry walls is defined two
ways masonry to resist all shear or steel to resist all shear. Height-to-
width ratio (h/d) also effects shear strength. Narrow walls have less
strength than long walls. Allowable shear stress depends on the ratio
M/Vd which may be expressed as h/d ratio (M/Vd = Vh/Vd = h/d).
1 Allowable shear stress Fv (assuming masonry resists all shear):
For h/d < 1
Fv = 1/3 (4-h/d) (ƒ’m)1/2 Fv (max) = 80 - 45h/d psi *
For h/d => 1
Fv = 1.0 (ƒ’m)1/2 Fv (max) = 35 psi *
2 Allowable shear stress Fv (assuming reinforcing resists all shear):
For h/d <1
Fv = 1/2 (4-h/d) (ƒ’m)1/2 Fv (max) = 120 - 45h/d psi *
6
For h/d =>1
*
3
Fv = 1.5 (ƒ’m)1/2
0 - 2 0 0
Cantilever wall, free to bend in single curvature
Fv (max) = 75 psi *
9 9
Wall with fixed support bends with inflection point at mid-height
1
erle
d Width of wall or wall element
h Height of wall or wall element
c h i
ƒ’m Specified masonry compressive strength (ksi)
GS
Fv Allowable shear stress (psi and MPa)
M Bending moment M = Vh or M = Vh/2
G
V Shear force
r i g h t Bar spacing s = Av Fs / bFv

Av= bar area, Fs= allowable bar stress, b= wall width, Fv= allowable masonry shear stress
Copy Example: CMU shear wall design

Assume: h=8’, 8” (nominal), ƒ’m=2 ksi, # 4 bars, Fs=24 ksi, no inspection, design masonry
to resist all shear of V = 6,000 #. Try wall length d = 4’, h/d=2
Allowable shear stress from graph 1
Bar spacing s = Av Fs / bFv = 0.2x24,000 / (7.625x17) = 37”
Fv = 17 psi
use s = 32”
Note: bar spacing rounded down to 8” CMU module
Wall shear capacity V = Fv (wall area) = 17x7.625x48” V = 6,222 #
Example: CMU shear wall analysis
Assume: same wall as above, but inspected and with reinforcing to resist all shear
Allowable shear stress from graph 2 Fv = 58 psi
Bar spacing s = Av Fs / bFv = 0.2x24,000 / (7.625x58) = 10.8” use s = 8”
Wall shear capacity V = Fv (wall area) = 58x7.625x48” V = 21,228 #
Note: rebars at 8” vs. 32” and inspection increase capacity from 6 k to 21 k

Specified Compressive strength ƒ’m for masonry is defined by the strength of masonry
units and mortars type M, S, N, with values from ƒ’m = 1,500 to 4,000 psi.
Allowable compressive stress Fa for masonry with special inspection with or without
grouting is 25% of the specified strength ƒ’m by the working stress method; reduced for
slenderness (shown in graph 1) as followings:
Fa = 0.25 ƒ’m [1-(h’/140r)2] * for h’/r ≤ 99 (h’/t ≤ 29)
Fa = 0.25 ƒ’m (70r/h’)2] * for h’/r > 99 (h’/t > 29)
For reinforced masonry columns the allowable compressive force Pa is:
Pa = (0.25ƒ’m A e + 0.65 A s Fsc) [1-(h’/140 r)2] * for h’/r ≤ 99 (h’/t ≤ 29)**
Pa = (0.25ƒ’m A e + 0.65 A s Fsc) (70 r/h’)2 * for h’/r > 99 (h’/t > 29)**
A e = Area of masonry (net area for un-grouted masonry)
A s = Area of steel reinforcement
F sc = Allowable compressive stress of steel reinforcement
6
h’= effective height of wall
- 2 0
thickness t, where r = (I / A)1/2 = 0.289 t
0 0
r = radius of gyration; for convenience, graph 1 substitutes radius of gyration r by
9
* Allowable stresses and loads are one half without special inspection
1 9
** For non-square columns the smaller dimension governs slenderness
erle
1 Slenderness reduction for allowable compressive stress
i
2 Masonry wall or column with pin support at both ends
GS c h3 Masonry wall or column with one fixed support

4 Masonry wall or column with two fixed supports
5 Masonry wall or column freestanding
i g h t G A Reduction factor for slenderness h’/t

B Slenderness vs. stress reduction curve
y r h Height of wall or column
Cop
h’ Effective height, adjusted for support type
t Wall thickness
Example: CMU wall
Assume: h=15’, both ends fixed, h’=0.6x15=9’, 8” CMU, t = 7.625”, f’m = 2000 psi
Find allowable stress Fa
Slenderness h’/t = 9’x12”/7.625 = 9.4
Slenderness reduction (from graph 1) A = 0.94
Fa = 0.25 f’m A = 0.25 x 2000 x 0.94 Fa = 470 psi
Example: brick column
Assume: brick column, 20”x24”, h = 30’, pin supports, f’m = 2,5 ksi, with 6 #8 steel bars,
grade 60, Fa = 60x0.4 = 24 ksi. Find allowable load P
Slenderness h’/t = 30’x12”/20” = 18, slenderness reduction (from graph 1), A = 0.81
P = (0.25ƒ’mA e+0.65A sFsc)(0.81)
P = (0.25x2.5x20x24+0.65x6x0.44x24)(0.81) P = 276 k

Balanced beam
Concrete Strength Design (LRFD) A convenient reference is the balanced beam whith steel reinforcing that reaches yield
Concrete strenth design is based on ultimatre concrete strenght, reduced by the strength simultaneously with concrete. However, actual reinforcimg should be less to
reduction factor φ, similar to LRFD. At ultimte stress, concrete yields, forming a parabolic assure ductile bahavior (steel yields before brittle concrete failure). Considring similar
stress block. But srenght design for rectangular beams assumes a rectangular stress triangles, balanced reinforcing ρb is derived, assuming Es = 29,000 ksi:
block wich gives similar results demonstrated by tests. Like masonry, concrete is strong
in compression but very weak in tension. Hence steel reinforcing is used to resist tenion. cb / 0.003 = d / (0.003 + fy/Es)
0.003 87
1 = Ultimate bending stress cb = (d) cb = (d)
0.003 + f7 /29, 000 87 + fy
2 = Bending stress assumed in strength design
3 = Strain of balanced beam For equilibrium (ΣH = 0, C = T) 0.85f’c β1 cb b = As fy
c = distance of neutral axis from top Thus cb = As fy / (0.85 f’c β1 b)
d = effective depth (from top of beam to centroid of steel) ρ b bdfy ρ df
Since As = bdρb 87
h = depth of beam cb = = b y = (d)
85f'c β 1b 85f'c β 1 87 + fy
6
f’c = specified concrete compressive strength
0
Solving for balanced ρb 0.85f' c β 1 ⎛ 87 ⎞
0
As = cross section area of steel As = bdρ ρb = ⎜⎜ ⎟⎟
- 2
fy ⎝ 87 + fy ⎠
0
fy = steel yield strength
a = depth of concrete stress block
β1 = 0.85 for f’c ≤ 4 ksi, reduced 0.5 per 1 ksi > 4 ksi, min β1 = 0.65
a = c β1
Thus
1 9
For equilibrium (ΣH = 0, C = T)
9
0.85 f’c ab = As fy
a = As fy / (0.85 f’c b)
erle
Z = resistant moment lever arm Z = d - a/2 For moment equilibrium (ΣM = 0) M = (C or T ) (d-a/2) = As fy (d-a/2)
i
C = concrete compression C = 0.85 f’c ab Substituting a and As = ρbd and (0.59 = (1/0.85)/2) and rearranging yields
T = steel tension
ρ = Percentage of reinforcement
G S c h
T = As fy
ρ = As / bd
⎛ f ⎞
M = bd2ρfy ⎜ 1- 0.59ρ y' ⎟
⎜
⎝ fc ⎟⎠
i g ht G Provides the nominal moment

Where R = Resistance factor ⎛
M = Rbd2
fy
R = ρfy ⎜⎜ 1- 0.59ρ '
⎞
r
⎟⎟
opy
⎝ fc ⎠
The nominal design moment is adjusted by a reduction factor φ = 0.9 φMn = M
C Reduction factors
Bending
Shear and torsion
Compression (spiral reinforcing)
φ = 0.90
φ = 0.85
φ = 0.75
Compression (tied reinforcing) and bearing φ = 0.70
Reinforcing ratio limits ρ
Minimum ρ = 0.2 ksi / fy
Recommended ρ = 0.18 f’c / fy
Maximum (75% of balanced reinforcing) ρ = 0.75 ρb
Minimum Resistance factor (at min. ρ = 0.2 ksi / fy) R = 0.192
Note: Balancd reinforcing implies steel and concrete provide equal (balanced) strenght
Less steel provides ductile steel behavior, rathter than britttle concrete failure.

Design graph
The design graph shows ρ-factors on the X-axis and R-factors on the Y-axis. The graph
lines extend from minimum to maximum ρ-factors with recommended values marked with
an ^. The following examples demonstrate use of the graph:
Example: beam design
Assume: simply supported beam; L = 16’; f’c = 3 ksi; Fy = 60 ksi
Factored Dead load = 1.4 x 943 plf / 1000 D = 1.32 klf
Factored Live load = 1.7 x 400 plf / 1000 L = 0.68 klf
w = 2.00 klf
Bending moment M = w L2/8 = 2 klf (16)2/8 M = 64 k’
Mn = M/φ = 64k’x12”/ 0.9 MR = 853 k”
6
Recommended depth (from table below) h = 16’(12”)/16 h = 12”
Effective depth (d = h-3” for bar+cover) d = 12-3
Recommended R-factor (from graph)
Beam width b = Mn/(Rd2) = 853/(0.483x92)
0 - 2 0 0 d = 9”
R = 0.483 ksi
b = 22”
9 9
Recommended reinforcement ratio (from graph)
1
ρ = 0.009
erle
Bar cross section
As = ρbd = (0.009)(22)(9) = 1.78 in2 Use 6 # 5 bars, As = 6x0.31 =1.86 in2
c h i
Example: alternate beam design
GS
Assume: the same beam to fit an 8” CMU wall, R = 0.483, ρ = 0.009
Effective depth
i g h t G d = [Mn/(R b)]1/2 = [853/(0.483x8”)]1/2

Rebar cross section
d = 15”
y r As = ρbd = (0.009)(8x15) = 1.08 in2 Try 3 # 6 bars, As = 3x0.44 = 1.32 in2
Cop
Check width: 3 bars+2 spaces+stirrups+cover = 3x6/8+2+1+3 = 8.25 > 8 not OK
Use 2 # 7 bars, As = 2x0.6 As =1.2 in2
Check width = 2x7/8+1+1+3 = 6.75 6.75 < 8, OK
Minimum depths h of beams and slabs unless deflections are computed (L = span)
Support type Beams & ribs One-way slabs
1 Simply supported L/16 L/20
2 One end continues L/18 L/24
3 Both ends continues L/21 L/28
4 Cantilever L/8 L/10
Bar diameters Cross-section areas
#5 5/8 0.625 15.9 0.31 200
#6 6/8 0.750 19.1 0.44 284
#7 7/8 0.875 22.2 0.60 387

Shear reinforcement
Shear reinforcement in bending members is required if the factured shear Vu exceeds
the shear capacity of concrete Vc,except for:
• Slabs and footings
• Concrete joists
• Beams of <10” depth
• Beams with Vu < φVc/2
Concrete shear capacity Vc = fc' bd
Subject to maximum shear stress Fv = 100 psi
Shear reinforcing is usually provided by vertical stirrups
Spacing s = Avfvd / Vs
Av = total cross section area of stirrups (usually 2 bars per stirrup)
6
Fy = yield stress of stirrups
Vs = shear resisted by stirrups
0 - 2 0
d = effective depth (top of beam to steel rebars)
0
9
Maximum spacing s = d/2
rle 1 9
Shear resisted by steel Vs = Vu/φ-Vc
i e
The maximum shear may be taken a distance d from supports
G Sc hReinforcing is required where

A Tensile reinforcement
V > Vc/2
G
b Beam width
ri g h t C
d
Shear reinforcing
Effective beam depth
Co p y E
s
Example
Distance from support requiring stirrups (at Vc/2)
Stirrup spacing
Design a simply supported beam, assume: f’c = 3 ksi, Fy = 60 ksi, L = 20’, b = 10”, d = 12”
Factored DL+LL w = 4 klf
Concrete shear capacity Vc = φ fc' bd = 0.85 4x10x12 Vc = 18 k
Maximum factored shear Vu = 4 klf x20’/2 Vu = 40 k
Shear at d from support Vu = 40-(4x12/12) Vu = 36 k
Shear resisted by steel Vs = Vu/φ-VC = 36 / 0.85-22 Vs = 20k
Try # 4 srirrups (As = 2x0,2 As = 0.4 in2
Spacing s=Avfvd / Vs = 0.4x60x12/20 s = 14”
Check max s = d/2 = 12/2 = 6” Use s = 6”
Distance stirrups needed E = (L/2/Vu)(Vu-Vc/2) = (10/40)(40-9) E =7.75’

Continuous and fixed-end beams
Concrete beams may continue over more than two supports, have moment resistant
(fixed-end) supports , or both; all of which cause negative support moments that reduce
positive mid-span moments and, therefore, require less depth than simply supported
beams. Fixed-end supports are usually in beams of moment resisting frames.
1 Simply supported beam (determinate)
2 Continuous beam over 3 supports
3 Continuous beam over 4 supports
4 Cantilever beam
5 Load diagram for beam 9
6 Shear diagram
7 Bending diagram
8
9
Deflection diagram
Beam reinforcement
- 2 0 0 6
A
B
9 9 0
Inflection points of zero bending moment, change from negative to positive bending
Top bars at negative bending and convex deflection
1
erle
C Hook at bar end of maximum stress anchors bar to concrete
D Bottom bars at positive bending and concave deflection
c h
E
i Stirrups resist shear stress, with increased spacing toward mid-span of zero shear
GS
i g h t G 1 Simply supported
2 One end continues
L/16
L/18
L/20
L/24
y r 3 Both ends continues L/21 L/28
Cop
4 Cantilever L/8 L/10
Reinforcement of continuous and fixed end beams follows shear and bending diagrams.
Shear is similar to simply supported beams. Moment distribution varies from positive at
mid-span to negative at supports and fixed-ends; causing convex deflections at supports
and concave at mid-span with change at inflection points. Reinforcing correlates with the
bending diagram: bottom bars at positive bending and top bars at negative bending; both
extending somewhat beyond the inflection points to account for variable live loads.
Note:
The reader is referred to books on reinforced concrete design (Spiegel, 1992) for issues
beyond the scope of this book, such as design of T-beams, beams with compression
reinforcement, bond length of bars, combined axial and compressive stress, etc.

T-Beam
Floor slabs are usually poured together with beams. This provides to combine slab and
beam as T-beam, with part of the slab acting as compressive flange and the beam acting
as stem or web. Reinforcement at the beam bottom resists tensile stress for positive
bending. For negative bending of continuous beams the tensile reinforcement must be
on top and the beam resists compression, without benefit of the wider slab. T-beams
are, therefore, most efficient as simply supported beams with positive bending only.
Shear resistance is limited to the cross section of the web or the area defined by the
width of the web and the effective depth of the beam. The flange width provided by the
slab is limited to 1/4 of the span, 16 times the slab thickness plus width of the beam, or
beam spacing, whichever is less. Depending on the ratio of reinforcement to
compressive area, the neutral axis of T-beams may be below or within the slab thickness.
6
For schematic design the resisting lever-arm may be estimated as the distance between
center of the slab and center of the reinforcement.
0 - 2 0 0
9
1 T-beam with compression zone depth a > slab thickness t
2
9
T-beam with compression zone depth a = slab thickness t
1
erle
3 T-beam with compression zone depth a < slab thickness t
4 T-beam with compression web due to negative bending
c ha
i Depth of compression zone
GS
b Width of compression flange, limited by ACI code to the lesser of:
● 1/4 beam span
i g h t G ● 16 times slab thickness plus web width

● beam spacing e from center to center
y r
Cop
d Effective depth (distance from reinforcement to compression zone edge)
Z Lever arm of internal resisting moment
(Distance from reinforcement to compression zone center)
NA Neutral Axis
t Slab thickness

One-way slabs
With reinforcing in only one direction, one-way slabs need reinforcing for temperature
variation and shrinkage perpendicular to the main reinforcing. As percentage of slab
cross section area, temperature reinforcing must be at least:
As = 0.20% for grade 40 and 50 steel
As = 0.18% for grade 60 steel
Simply supported L/16 L/20
One end continues L/18 L/24
Both ends continues L/21 L/28
Cantilever L/8 L/10
Bar diameters
Size in in mm
- 2 0 0 6 Cross-section areas
in2 mm2
0
#4 4/8 0.500 12.7 0.20 129
#5
#6
5/8
1
6/8
9
0.625
0.750
9
15.9
19.1
0.31
0.44
200
284
erle
Example: One-way slab design
c h i
Assume: simply supported slab, L = 16’; f’c = 3 ksi; Fy = 40 ks, Design a 1’ wide strip
GS
Slab depth h = L / 20 = 16’x12”/20 h = 9.6”
Dead load = 150pcf x9.6”/12” = 120 psf+20 psf partitions+14 psf misc. DL = 154 psf
G
Factored loads: (1.4x154 psf DL+1.7x50 psf LL)/1000 w = 0.3 klf
r i g h t Moment M = wL2/8 = 0.3x162/8

Required resisting moment (k”)
M = 9.6 k’
Copy Mn = 12”M/φ = 12”x9.6k’/0.9

Effective depth d = h – bar/2 - cover
d = 9.6 – 0,75/2 - 0.75
Resistance factor R = Mn / bd2 = 128 / (12x8.52)
Mn = 128 k”
d = 8.5
R = 0.148 < 0.192
For min. R = 0.192 min. steel ratio ρ = 0.2 ksi / fy ρ = 0.005
Bar area As = ρbd = 0.005x12x8.5 As = 0.51 in2
Use # 6 bars, As = 0.44 in2 per bar
Bar spacing s = 12”x 0.44 / 0.51 s = 10.3”
Temperature reinforcing
As = 0.002bd = 0.002x12x8.5 As = 0.204 in2
Try #6 bars
Bar spacing: s = 12”x 0.44 / 0.204 s = 25.9”
Check s vs. ACI spacing limits: 4” < s < 18” < 5h 25.9 > 18, not OK
Use # 4bars
Bar spacing: s = 12”x0.20 / 0.204 s = 11.8”

Two-way slabs and plates
Two-way systems should have about equal spans both ways. Double spans increase
deflection 16 times (4th power of span). They may be thick plates or thin slabs with drop
panels at posts to resist shear. They can be designed by Direct Design Method,
assuming : 1) at least 3 spans; 2) span ratios ≤ 2:1; 3) adjacent spans differ < 1:1.3; 4)
post offsets < 1.1L; 5) uniform load L ≤ 2 D. Bending M = M0 x coefficient (at left), where:
M0 = w L2/8 L = span; w = 1.4D+1.7L
1 Column strips and middle strips of typical slab
2 Moment coefficients for slab and plate with simply supported end span
3 Moment coefficients for slab and plate supported directly on columns
4 Moment coefficients for slab and plate with edge beam
5 Moment coefficients for slab and plate with end span integral with wall
6
A Column strip (slab/column moment distribution is not considered)
B
C
Middle strip
End support (negative moment)
0 - 2 0 0
9
D End span (positive moment)
E
9
First interior support (negative moment)
1
erle
F Interior span (positive moment)
G Interior support (negative moment)
c h
H
i Column strip moment coefficients
GS
I Middle strip moment coefficients
Minimum depths h of two-way slabs unless deflections are computed (L = span)
i g h t G Support type
Plate without edge beams
ƒy = 40 ksi (276 MPa)
L/33
ƒy = 60 ksi (414 MPa)
L/30
y r Plate with edge beams L/36 L/33
Cop
Slab with drop panels without edge beams L/36 L/33
Slab with drop panels and edge beams L/40 L/36
Example: Bending moment for slab 2 with simple end span
Assume: Span L = 20’, factored dead + live load w = 0.3 klf
Column strip moments M0 = wL2/8 = 0.3x202/8 Mo = 15 k’
End span M = +0.38x15 M = +5.7 k’
First interior post M = -0.56x15 M = -8.8 k’
Interior span M = +0.21x15 M = +3.15 k’
2nd interior post M = -0.49x15 M = -7.35 k’
Middle strip
End span M = +0.25x15 M = +3.75 k’
First interior post M = -0.19x15 M = -2.85 k’
Interior span M = +0.14x15 M = +2.10 k’
2nd interior post M = -0.16x15 M = -2.40 k’
Reinforcing is similar to one-way slabs, but two ways, without temperature reinforcing.

Column
Concrete columns may have square, rectangular, round, or of other cross section with
tied or spiral rebars. Compression bars shall be min. No. 5 (16 mm) or greater. The ACI
code limits reinforcement to 1% min. and 8% max. as percentage of column cross
section area; but 4% is recommended to prevent rebar crowding. Bars and ties require
1.5 in (38 mm) concrete cover for fire and corrosion protection. Columns of width / height
ratios of less than 13 are designed short columns without considering buckling.
Tied columns must need at least four compression bars held in place by ties: No. 3 ties
for compression bars up to No. 10 and No. 4 ties for larger ones. Ties shall secure all
corner bars and at least every second bar between corners. Unsecured bars shall be not
more than 6 in (15 cm) from a secured bar. Tie hooks shall be 135º. Tied columns are
more common, provide somewhat less strength, are less expensive than spiral columns,
and adapt easier to cross-, T-, U-, and L-shaped columns. Maximum tie spacing shall be
2 0 0 6
16 bar diameters, 48 tie diameters, or the least column dimension, whichever is less. For
seismic design maximum tie spacing shall be 1/2 the least column dimension near beam
-
0
intersections.
1 9 9
Spiral columns must have at least five or more vertical compression bars in circular
erle
configuration held in place by a continuous circular spiral of about 1/4 in diameter. Spiral
columns are usually cylindrical, but spiral reinforcing may also be used for square
c h i
columns. Spiral columns are about 14% stronger than tied columns of equal cross
GS
section area because spirals confine the concrete and rebars better under high stress.
Spiral spacing ranges from min. 1 in (25 mm) to max. 3 in (76 mm).
i g h t G 1
2
Square column with tied reinforcement of minimum 4 bars
Round column with spiral reinforcement of minimum 5 bars
y r 3 Square column with spiral reinforcement
Cop
4 Rectangular column with tied reinforcement
5 Square column with tied reinforcement of 8 bars
6 Round column with 2-ring spiral reinforcement
7 Square column with tied reinforcement of 16 bars
8 Cross-shaped column with tied reinforcement
9 L-shaped column with tied reinforcement

Column design
The strength of concrete columns is defined by concrete strength, grade, amount, and
type of steel reinforcing. The theoretical strength without eccentricity is:
Po = 0.85 f’c(Ag – As) + fyAs
Ag = column cross section area
As = area of steel reinforcing
f’c= specified concrete compressive strength
fy = yield strength of steel reinforcing
However, concrete columns may be subject to eccentric load or bending moments from
beams. Therefore, ACI assumes an implied eccentricity by reduction factors of 0,85 for
spiral columns and 0.80 for tied columns. In addition, strength is reduced by Ф = 0.75 for
spiral columns and Ф = 0.70 for tied columns. Thus, ACI defines column strength as:
Spiral columns ФP = 0.85Ф [0.85f’c (Ag – As) + fy As]
ФP = 0.80Ф [0.85f’c (Ag – As) + fy As]
6
Tied columns
For spiral columns
0 - 2 0 0
For convenient schematic design formulas for stress used in the graphs are:
F = 0.75x0.85 [0.85f’c (1 - ρ) + fy ρ]
F = 0.70x0.80 [0.85f’c (1 - ρ) + fy ρ]
9
For tied columns
1
Design graphs
9
erle
The design graphs for tied columns at left and spiral columns on the next page are based
on the above equations. Their use is described by examples.
c h i
Example: Tied columns, 3-story
GS
Assume: Lateral load resisted by shear walls, design for gravity load only
Tributary area 30’x30’, DL = 175 psf, LL 50 psf, fy = 60 ksi
G
Factored load w = 1.4x175+ 1.7x50 w = 330 psf
r i g h t Ground floor (use f’c = 5 ksi, 4% steel) F = 3.6 ksi

P = 3x30’x30’x330/1000 P = 891k
Copy A = P/F = 891/3.6

Column size, A1/2 = 2481/2 = 15.7”
Steel area As = 0.04x162 = 10.2 in2
First floor (use the same column with f’c = 3 ksi)
A = 248 in2
Use 16”x16”
Use 18 # 7 bars, As = 10.8 in2
P = 2x30’x30’x330/2000 P = 594 k
F = P/A = 594/162 = 2.32 ksi Use 2.8% steel
Steel area As = 0.028x162 = 7.2 in2 Use 12 # 7 bars, As = 7.2 in2
Second floor (use same column with fy = 40 ksi)
P =1x30x30x330/1000 P = 297 k
F = P/A = 297/162 =1.16 ksi Use 1% steel
Rebar diameters Cross-section areas
#7 7/8 0.875 22.2 0.60 387
#8 8/8 1.000 25.4 0.79 510

Example: Spiral columns, 3-story
Assume: The same project as above but with spiral columns
Tributary area 30’x30’, fy = 60 ksi, DL = 175 psf, LL 50 psf.
Factored load w = 1.4x175+ 1.7x50 w = 330 psf
Ground floor (use f’’c = 5 ksi, 5% steel) F = 4.5 ksi
P = 3x30’x30’x330/1000 P = 891k
A = P/F = 891/4.5 A = 198 in2
Column size, 2(A/π)1/2 = 2(198/π)1/2 = 15.8” Use φ16”
Column cross section area A = πr2 = π(16/2)2 A = 201 in2
First floor (use the same column with f’c = 3 ksi)
P = 2x30’x30’x330/2000 P = 594 k
F = P/A = 594/198 = 2.98 ksi Use 3.8% steel
6
Second floor (use the same column with fy = 40 ksi)
P =1x30x30x330/1000
0 - 2 0 0 P = 297 k
9
F = P/A = 297/198 = 1.5 ksi Use 1% steel
1 9
c h i erle
h t G GS
y r i g
Cop

Lateral loads, acting primarily horizontally, include:
9 •
•
•
Wind load
Seismic load
Earth pressure on retaining walls (not included in this book)
Wind and earthquakes are the most devastating forces of nature:
Hurricane Andrew 1992, with gusts of 170 mph, devastated 300 square miles, left
300,000 homeless, caused about $ 25 billion damage, and damaged 100,000 homes
La t e ra l Forc e De sign The 1976 Tangshan Earthquake (magnitude 7.8), obliterating the city in northeast China
and killing over 240,000 people, was the most devastating earthquake of the 20th century.
6
Swiss Re reported 2003 world wide losses:
• 60,000 people killed
• Over two thirds earthquake victims
0 - 2 0 0
9
• $70 billion economic losses
1 9
c h i erle
IBC table 1604.5. Importance Classification excerpt
GS
importance
importance
importance
Use Group
Category
Seismic
Seismic
Nature of Occupancy
factor
factor
factor
Snow
Wind
r i g h t
y
I I Low hazard structures: 1 0.8 0.87
Cop
Agriculture, temporary, minor storage
II I Structures not in categories I, III, IV 1 1 1

III II Structures such as: 1.25 1.1 1.15
Occupancy >300 people per area
Elementary schools >250 students
Colleges >500 students
Occupancy >5000
IV III Essential facilities, such as: 1.5 1.2 1.15

Hospitals, polices and fire stations
9-1 DESIGN METHODS Lateral Force Design

Wind load
1 Wind load on gabled building
2 Wind load on dome or vault
3 Protected buildings inside a city
4 Exposed tall building inside a city
5 Wind flow around and above exposed building
6 Wind speed amplified by building configuration
Wind channeled between buildings causes a Venturi effect of increased wind speed. Air
movement through buildings causes internal pressure that affects curtain walls and
cladding design. Internal pressure has a balloon-like effect, acting outward if the wind
enters primarily on the windward side. Openings on leeward or side walls cause inward
pressure. In tall buildings with fixed curtain wall the difference between outside wind
pressure and interior pressure causes air movement from high pressure to low pressure.
2 0 0 6
This causes air infiltration on the windward side and outflow on the leeward side. In high-
rise buildings, warm air moving from lower to upper levels causes pressures at top levels
-
0
on the leeward face and negative suction on lower levels. Wind pressure is based on the
9 9
equation developed by Daniel Bernoulli (1700-1782). For steady air flow of velocity V,
1
erle
the velocity pressure, q, on a rigid body is
q = pV2 / 2
c h i
p = air density (air weight divided by the acceleration of gravity g = 32.2 ft/sec2)
GS
Air of 15°C at sea level weighs 0.0765 Ib/ft3, which yields:
i g h t G q = 0.00256V2 (q in psf)
r
The American National Standards Institute (ANSI) Minimum design loads for buildings
Copy and other structures (ANSI A58.1 - 1982), converted dynamic pressure to velocity
pressure qz (psf) at height z as
qz = 0.00256 Kz (I V)2
Kz = 2.58(z/zg)2/a (for buildings of 15 ft or higher)
a = Power coefficient (see exposures A – D below)
Z = Height above ground
Zg = Height at which ground friction no longer effects the wind speed
(see exposures A - D below)
I = Importance factor (see IBC table 1604.5)
ANSI A58.1 defined exposures A, B, C, D (IBC uses B, C, D only):
Exposure A Large city centers a = 3.0, Zg = 1500 ft
Exposure B Urban and suburban areas, wooded areas a = 4.5, Zg = 1200 ft
Exposure C Flat, open country with minimal obstructions a = 7.0, Zg = 900 ft
Exposure D Flat, unobstructed coastal areas a = 10.0, Zg = 700 ft

Exposure C wind pressure for 10' to 100' height and 85 to 150 mph wind speed Example: Wood shear walls
45.0
40.0 Assume: 66’x120’x27’ high, 3 shear walls, L=3x30’=90’, wind speed 90 mph, exposure C,
Windward pressure (psf)
35.0 Importance factor I = 1, gust factor G = 0.85 (ASCE 7, 6.5.8 for rigid structures > 1 Hz)
30.0
For each level in width direction find: wind pressure P, force F, shear V, shear wall type
25.0
20.0
Interior pressure (from graph for h = 30’) p = 3.1 psf
15.0 Leeward suction (from graph for h = 30’ P= 7.3 psf
10.0
5.0
Level 3 (h = 29 – use 30’ pressure)
0.0 Wind pressure (windward + leeward + interior)
10 20 30 40 50 60 70 80 90 100
p = 11.7+7.3 + 3.1 p = 22.1 psf
85 9.1 9.6 10.5 11.1 11.7 12.1 12.5 12.9 13.3 13.5
90 10.2 10.8 11.7 12.5 13.1 13.5 14.0 14.5 14.9 15.1 Force F = 22.1x120x10’/2 F = 13,260 #
100 12.6 13.3 14.5 15.4 16.1 16.7 17.3 17.9 18.3 18.6 Shear V = F V = 13,260 #
110 15.2 16.1 17.5 18.6 19.5 20.2 20.9 21.7 22.2 22.6
6
Required wall strength = 13,260 / 90’ = 147 plf; use 5/16”, 6d at 6” 200 > 147
0
120 18.1 19.2 20.9 22.2 23.2 24.1 24.9 25.8 26.4 26.8
0
130 21.3 22.5 24.5 26.0 27.3 28.3 29.3 30.3 31.0 31.5
2
Level 2 (h = 19’ – use 20’ pressure)
-
140 24.7 26.1 28.4 30.2 31.6 32.8 33.9 35.1 36.0 36.5
0
p = 10.8+7.3+3.1 p = 21.2 psf
9
150 28.3 30.0 32.6 34.6 36.3 37.6 39.0 40.3 41.3 41.9
9
Windward pressure (psf) Force F = 21.2 x 120 x 10’ F = 25,440 #
85 5.7 6.0 6.5 6.9 7.3 7.6 7.8 8.1 8.3 8.4
erle 1
Shear V = 13,260 + 25,440
Required wall strength = 38,700 / 90’ = 430 plf; use 15/32”, 8d at 4”
V = 38,700 #
430 = 430
i
90 6.4 6.7 7.3 7.8 8.2 8.5 8.8 9.1 9.3 9.4
c h
100 7.9 8.3 9.1 9.6 10.1 10.5 10.8 11.2 11.5 11.7
Level 1 (h = 9’ – use 10’ pressure’)
GS
110 9.5 10.1 11.0 11.6 12.2 12.6 13.1 13.5 13.9 14.1
120 11.3 12.0 13.1 13.8 14.5 15.0 15.6 16.1 16.5 16.8 p = 10.2+7.3+3.1 p = 20.6 psf
13.3 14.1 15.3 16.3 17.0 17.7 18.3 18.9 19.4 19.7
130
Force F = 23.7x120x10’ F = 24,720 #
140
150
15.4
17.7
16.3
18.7
17.8
20.4
i g h t
18.9
21.6
G 19.8
22.7
20.5
23.5
21.2
24.3
21.9
25.2
22.5
25.8
22.8
26.2 Shear V = 38,700 + 24,720
Required strength = 63,420 / 90’ = 705 plf; use 15/32”, 8d at 2”
V = 63.420 #
730 > 705
r
Leeward pressure (psf)
85
90
100
110
120 Cop
2.4
2.7
3.3
4.0
4.8
2.5
2.9
3.5
4.3
5.1
y
2.8
3.1
3.8
4.6
5.5
2.9
3.3
4.1
4.9
5.9
3.1
3.5
4.3
5.2
6.1
3.2
3.6
4.4
5.4
6.4
3.3
3.7
4.6
5.5
6.6
3.4
3.8
4.7
5.7
6.8
3.5
3.9
4.9
5.9
7.0
3.6
4.0
4.9
6.0
7.1
Note:
The results are very similar to 9.1 with less computation
See Appendix C for exposure B and D graphs
IBC table 2306.4.1 excerpts

130 5.6 6.0 6.5 6.9 7.2 7.5 7.7 8.0 8.2 8.3
6.5 6.9 7.5 8.0 8.4 8.7 9.0 9.3 9.5 9.7
Allowable shear for wood panels with Douglas-Fir-Large or Southern Pine
140
150 7.5 7.9 8.6 9.2 9.6 10.0 10.3 10.7 10.9 11.1 Nail spacing at panel edge (inches)
Interior pressure (psf)
Panel Panel Nail 6 4 3 2
Nail size
grade thickness penetration Allowable shear (lbs / foot)
5/16 in 1 1/4 in 6d 200 300 390 510
Structural I
sheathing
3/8 in 1 3/8 in 8d 230 360 460 610

7/16 in 1 3/8 in 8d 255 395 505 670
1 3/8 in 8d 280 430 550 730
15/32 in
1 1/2 in 10d 340 510 665 870
* Requires 3 x framing and staggered nailing

Exposure D wind pressure for 10' to 100' height and 85 to 150 mph wind speed
50.0
Example: CMU shear walls
45.0 Assume: Regular flat site
40.0 Office building: 6-story, 90’x90’x60’, 30’x30’ core
35.0 30’ CMU walls, 8” nominal (7.625”)
30.0
Shear wall length L = 2 x (30’– 6’ doors) = 48’
25.0
20.0
Shear walls resist all lateral load
15.0 Roof fabric canopy, 50’x50’x10’, gust factor G = 1.8
10.0 Wind speed V = 100 mph
5.0 Exposure D
0.0
10 20 30 40 50 60 70 80 90 100 Importance factor I = 1
85 11.0 11.5 12.4 13.0 13.6 14.0 14.3 14.8 15.0 15.3
90 12.3 12.9 13.9 14.6 15.2 15.7 16.1 16.5 16.8 17.1
100 15.2 16.0 17.2 18.1 18.8 19.4 19.8 20.4 20.7 21.2
Interior pressure (assume conservative opening height h = 60’) p = 5.1 psf
6
18.4 19.3 20.8 21.8 22.7 23.5 24.0 24.7 25.1 25.6 Leeward pressure (for h = 60’)
0
110
P = 12.1 psf + 5.1 psf p = 17.2 psf
0
120 21.9 23.0 24.7 26.0 27.1 27.9 28.6 29.4 29.8 30.5
- 2
130 25.8 27.0 29.0 30.5 31.8 32.8 33.5 34.5 35.0 35.8
Average windward pressure (h = 10 to 60’)
0
140 29.9 31.3 33.6 35.4 36.8 38.0 38.9 40.0 40.6 41.5
9
150 34.3 36.0 38.6 40.6 42.3 43.6 44.6 45.9 46.6 47.6 P = (15.2+16.0+17.2+18.1+18.8+19.4) / 6 p = 17.5 psf
9
Average combined wind pressure
Winward pressure (psf)
85 6.9 7.2 7.8 8.2 8.5 8.8 9.0 9.2 9.4
e
9.6
rle
P = 17.5+17.2+5.1
1
Roof canopy pressure P canopy = (12.1)(1.8)
P = 39.8 psf
P canopy = 21.8 psf
h i
90 7.7 8.1 8.7 9.1 9.5 9.8 10.0 10.3 10.5 10.7
Base shear
c
100 9.5 10.0 10.7 11.3 11.7 12.1 12.4 12.8 12.9 13.2
S
110 11.5 12.1 13.0 13.7 14.2 14.7 15.0 15.4 15.7 16.0 V = A P = 90’ x (60’-5’) x 39.8 + (50’x10’/2)x21.8 V = 202,460 #
G
120 13.7 14.4 15.4 16.2 16.9 17.4 17.8 18.4 18.6 19.0
130 16.1 16.9 18.1 19.1 19.8 20.5 20.9 21.6 21.9 22.3
Core shear stress
G
v =V/A = 202,460 / (48’x12”x7.625”) v = 46 psi
t
140 18.7 19.6 21.0 22.1 23.0 23.7 24.3 25.0 25.4 25.9
h
150 21.4 22.5 24.1 25.4 26.4 27.3 27.9 28.7 29.1 29.8
rig
Leeward pressure (psf) Note:
85
90
100
110
120
130 C o
2.9
3.3
4.0
4.9
5.8
6.8
p y 3.1
3.4
4.2
5.1
6.1
7.1
3.3
3.7
4.5
5.5
6.5
7.7
3.5
3.9
4.8
5.8
6.9
8.1
3.6
4.0
5.0
6.0
7.2
8.4
3.7
4.2
5.1
6.2
7.4
8.7
3.8
4.3
5.2
6.4
7.6
8.9
3.9
4.4
5.4
6.5
7.8
9.1
4.0
4.4
5.5
6.6
7.9
9.3
4.0
4.5
5.6
6.8
8.1
9.5
Wind on lower half of first floor, resisted by footing, has no effect on shear walls
140 7.9 8.3 8.9 9.4 9.7 10.1 10.3 10.6 10.7 11.0
150 9.1 9.5 10.2 10.8 11.2 11.5 11.8 12.2 12.3 12.6
Interior pressure (psf)

Seismic Design
Earthquakes are caused primarily by release of shear stress in seismic faults, such as
the San Andreas Fault, that separates the Pacific plate from the North American plate,
two of the plates that make up the earth’s crust according to the plate tectonics theory.
Plates move with respect to each other at rates of about 2-5 cm per year, building up
stress in the process. When stress exceeds the soil’s shear capacity, the plates slip and
cause earthquakes. The point of slippage is called the hypocenter or focus, the point on
the surface above is called the epicenter. Ground waves propagate in radial pattern from
the focus. The radial waves cause shaking somewhat more vertical above the focus and
more horizontal far away; yet irregular rock formations may deflect the ground waves in
random patterns. The Northridge earthquake of January 17, 1994 caused unusually
strong vertical acceleration because it occurred under the city.
0 0 6
Occasionally earthquakes may occur within plates rather than at the edges. This was the
- 2
case with a series of strong earthquakes in New Madrid, along the Mississippi River in
1 9
explosions, or similar man-made events
9 0
Missouri in 1811-1812. Earthquakes are also caused by volcanic eruptions, underground
erle
Buildings are shaken by ground waves, but their inertia tends to resists the movement
i
which causes lateral forces. The building mass (dead weight) and acceleration affects
GS c hthese forces. In response, structure height and stiffness, as well as soil type affect the
response of buildings to the acceleration. For example, seismic forces for concrete shear
walls (which are very stiff) are considered twice that of more flexible moment frames. As
i g h t G an analogy, the resilience of grass blades will prevent them from breaking in an
earthquake; but when frozen in winter they would break because of increased stiffness.
y r
Cop
The cyclical nature of earthquakes causes dynamic forces that are best determined by
dynamic analysis. However, given the complexity of dynamic analysis, many buildings of
regular shape and height limits, as defined by codes, may be analyzed by a static force
method, adapted from Newton’s law F= ma (Force = mass x acceleration).
1 Seismic wave propagation and fault rupture
2 Lateral slip fault
3 Thrust fault
4 Building overturn
5 Building shear
6 Bending of building (first mode)
7 Bending of building (higher mode)
E Epicenter
H Hypocenter

Basic concepts
Earthquake ground shaking generates forces on structures. Though these forces act in
all direction, the horizontal (lateral) forces are usually most critical. Seismic forces are
f = m a (Force = mass x acceleration )
m = mass (building dead load)
a = acceleration (Spectral Acceleration)
Note:
Spectral Acceleration approximates the acceleration of a building, as modeled by a
particle on a mass-less vertical rod of the same period of vibration as the building.
PGA (Peak Ground Acceleration) is experienced by a particle on the ground
Spectral acceleration
Acceleration Spectra (left)
- 2 0 0 6
0
Based on the 1971 San Fernando and other Earthquakes Seed (1976) developed
9 9
Acceleration Spectra to correlate time period (X-axis) with acceleration for four soil types.
1
Other studies by Hall, Hayashi, Kuribayashi, and Mohraz demonstrated similar results.
erle
Equivalent Lateral Force Analysis is based on Acceleration Spectra, abstracted as
i
Design Response Spectrum
GS c h
i g h t G
y r
Cop
Acceleration spectra for four soil types (by Seed)
Design Response Spectrum (left)

The IBC Design Response Spectrum correlate time period T and Spectral Acceleration,
defining three zones. Two critical zones are:
T < TS governs low-rise structures of short periods

T > TS governs tall structures of long periods
where
T = time period of structure (T ~ 0.1 sec. per story - or per ASCE 7 table 1615.1.1)
TS = SDS / SD1 (See the following graphs for SDS and SD1)
IBC Design Response Spectrum

Analysis steps
Define site class by geologist, or assume default site class D (IBC table 1615.1.1)
Define Mapped Spectral Accelerations SS and S1
For overview see USGS maps at left: 0.2 sec low-rise (top) 1 sec high-rise (bottom)
Enter Site coordinates at USGS web site:
http://eqdesign.cr.usgs.gov/html/lookup-2002-interp-D6.html
Enter latitude in the left box in decimal degrees (range: 24.6 to 50.0)
Enter negative longitude in the right box (range: -125.0 to -65.0)
Web output:
6
LOCATION 37.7795 Lat. -122.4195 Long.
10%PE in 50 yr. 2%PE in 50 yr.

0.2 sec SA 115.35
0 - 2 0 0
Interpolated Probabilistic Ground Motion(Spectral Acceleration SA) in %g, at the site are:
182.76 % → SS = 1.83 (for low-rise)
9
1.0 sec SA 53.08
1 9
92.41 % → S1 = 0.92 (for high-rise)
erle
Low-rise: T<TS (structures < 5 stories)
High-rise: T>TS (structures > 10 stories)
c h i TS = SDS/SD1 (For SDS and SD1 see graphs on following pages)
GS
Define base shear V (lateral force at base of structure)
V = CS W
G
W = Dead load (+ 25% storage live load + 20% flat roof snow load > 30 psf)
r i g h t CS= seismic coefficient - see sample graph at left (SS at top line)
For other structures:
0.50 Cop y
Cs factors for light fram ing with wood panels (R=6, I=1)
CS = I SDS / R
Need not exceed
CS = I SD1 / (TR)
I = Importance factor
R = R-factor
(for T < TS)
(for T > TS)

(IBC table 1604.5)
(IBC table 1617.6.2)
0.45
0.40 SDS and SD1 (See graphs on the following pages)
0.35 CS varies with spectral acceleration SS & S1 and type of structure
0.30
0.25
(defined on the following pages)
0.20 For example, in seismic areas:
0.15
E
CS ~ 3 % for tall steel frame structures
0.10
C
D
CS ~ 15 % for low-rise wood structures
0.05 B
0.00 A CS ~ 30 % for some low-rise masonry structures
0.25 0.5 0.75 1 1.25 1.5 2 3 4 5 6
A 0.02 0.03 0.05 0.06 0.08 0.10 0.13 0.19 0.25 0.32 0.38
W = w A (w = dead load, DL in psf, A = total gross floor area of building)
B 0.02 0.04 0.06 0.08 0.10 0.12 0.16 0.24 0.32 0.40 0.48 w varies with type of construction – for example:
C 0.02 0.05 0.07 0.08 0.10 0.12 0.16 0.24 0.32 0.40 0.48 w ~ 15 to 25 psf for wood structures
D 0.03 0.06 0.07 0.09 0.10 0.12 0.16 0.24 0.32 0.40 0.48
E 0.05 0.07 0.07 0.07 0.09 0.11 0.14 0.21 0.29 0.36 0.43 w ~ 70 to 100 psf for steel structures
Cs for site class A-E for 0.2 sec mapped spectral accelerations Ss (top line) w ~ 150 to 200 psf for concrete structures

IBC table 1615.1.1 Site class definitions excerpts Example: One-story residence, San Francisco
Site class Soil profile name Average shear velocity in top 100 ft (30 m) Assume: Light framing with plywood panels
A Hard rock Vs > 5000 ft/s (1500 m/s) 36’x40’x10’ high, DL = 25 psf, site class undefined, use default D, I = 1
B Rock Vs = 2500 to 5000 ft/s (760 to 1500 m/s) Enter site coordinates at USGS web site
http://eqdesign.cr.usgs.gov/html/lookup-2002-interp-D6.html
C Very dense soil & soft rock Vs = 1200 to 2500 ft/s (370 to 760 m/s)
Web site output
D Stiff soil Vs = 600 to 1200 ft/s (180 to 370 m/s)
0.2 sec Spectral Acceleration Ss = 1.85
E Soft soil Vs < 600 ft/s (180 m/s) Design Spectral Accelerations (see graph)
F Soil vulnerable to failure, very organic clay, high plasticity clay, etc. At Ss = 2.0 CS = 0.16
Interpolate CS at Ss = 1.85 (CS /1.85 = 0.16 / 2.0)
IBC table 1617.6.2 excerpt R-factor Height limits (ft), categories A-F CS = 1.85 x 0.16 / 2.0 CS =0.15
Bearing wall systems A B C D E F Building dead weight
Light framed walls with wood panels 6 NL NL 65 65 65 W = 25 psf x 36’x40’ W = 36,000#
6
Light framed walls with other panels 2 NL NL 35 NP NP Base Shear
Ordinary reinforced concrete walls
Special reinforced concrete walls
4
5
NL
NL
NL NP NP NP
NL 160 160 100
V = CS W = 0.15 x 36,000
0 - 2 0 0 V = 5,400 #
9
Ordinary reinforced masonry walls 2 NL 160 NP NP NP Example: Same residence in San Francisco on site class A
Special reinforced masonry walls 5 NL NL 160 160 100
1 9
CS factor (see graph)
erle
Building frame systems Interpolate CS at Ss = 1.85 (CS /1.85 = 0.13 / 2.0)
CS = 0.13x1.85/ 2.0 CS = 0.12
i
Ordinary steel concentric braced frames 5 NL NL 35 35 NP
h
Base shear V = CS W = 0.12 x 36,000 V = 4,320 #
c
Special steel concentric braced frames 6 NL NL 160 160 100
GS
Ordinary steel moment frames 3.5 NL NL NP NP NP
Special steel moment frames 8 NL NL NL NL NL Example: Same residence in Tucson
G
Site class D, SS = 0.329
r i g h t CS factor (see graph)

Interpolate for Ss = 0.329 (CS/0.329 = 0.06/05)
Cop y
Cs factors for light fram ing with wood panels (R=6, I=1)
CS = 0.329x0.06 / 0.5
Base shear V = CS W = 0.04 x 36,000
Example: Same residence in Tucson on site class A

CS = 0.04
V = 1,440 #
0.50 CS factor (see graph)

0.45 Interpolate for Ss = 0.329 (at Ss = 0.5 CS = 0.03)
0.40
0.35
CS = 0.329 x 0.03 / 0.5 CS = 0.02
0.30 Base shear V = CS W = 0.02 x 36,000 V = 720 #
0.25
0.20 Compare seismic factors
0.15
0.10
E Los Angeles site class D Cs = 0.15
D
0.05 B
C Los Angeles site class A Cs = 0.12
0.00 A Tucson site class D Cs = 0.04
0.25 0.5 0.75 1 1.25 1.5 2 3 4 5 6
A 0.02 0.03 0.05 0.06 0.08 0.10 0.13 0.19 0.25 0.32 0.38
Tucson site class A Cs = 0.02
B 0.02 0.04 0.06 0.08 0.10 0.12 0.16 0.24 0.32 0.40 0.48
C 0.02 0.05 0.07 0.08 0.10 0.12 0.16 0.24 0.32 0.40 0.48
D 0.03 0.06 0.07 0.09 0.10 0.12 0.16 0.24 0.32 0.40 0.48
E 0.05 0.07 0.07 0.07 0.09 0.11 0.14 0.21 0.29 0.36 0.43
Cs for site class A-E for 0.2 sec mapped spectral accelerations Ss (top line)
Vertical distribution
Seismic forces increase with building height since f = ma (force = mass x acceleration),
i.e., increased drift increases acceleration. Thus story forces Fx are story mass times
height above ground. For buildings with periods of 0.5 seconds or less the force increase
is considered liner. For tall buildings the story-force varies non-linear. Since all story
forces are resisted at the ground, each story must resist its own force plus all forces from
above. Thus shear per level increases from top to bottom. The overturn moment per
level is the sum of all forces above times their distance to the level considered.
1 Linear force increase for T ≤ 0.5 seconds
2 Non-linear force increase for T > 0.5
3 Distribution per level of force
Fx = force per level x
Vx = Shear per level x = sum all forces above
V2 = 3 k
V1 = 3 k + 2 k
- 2 0 0 6 V2 = 3 k
V1 = 5 k
0
V0 = 5 k + 1 k V0 = 6 k
9 9
Mx = overturn moment per level = sum of all forces above times level arm
1
Assuming 10’ story height:
erle
M2 = 3 k x10’ M2 = 30 k’
i
M1 = 3 k x20 + 2 k x 10’ M1 = 80 k’
GS c h M0 = 3 k x 30’ + 2 k x 20’ + 1 k x 10’

4 Overturn moment visualized
5 Force per level
M0 = 140 k’
h t G Fx =Cvx V
rig
n
C vx = w x hx / ∑ w i hik (vertical distribution factor)
C o p y i=1
W = total dead weight of level x
h = height of level x above ground
n = total number of stories
k = exponent related to structure period
k = 1 for T ≤ 0.5 seconds
k = 2 for T > 2.5 seconds
k = interpolated between T = 0.5 and 2.5
6 k Interpolation graph
7 Shear per level
n
Vx = ∑ Fi
i=x
8 Overturn moment per level
n
M x = ∑ F(h
i n - hi )
i=x

Horizontal diaphragms
Horizontal floor and roof diaphragms transfer lateral load to walls and other supporting
elements. The amount each wall assumes depends if diaphragms are flexible or rigid.
1 Flexible diaphragm
Floors and roofs with plywood sheathing are usually flexible; they transfer load,
similar to simple beams, in proportion to the tributary area of each wall
Wall reactions R are computed based on tributary area of each wall
Required shear flow q (wall capacity)
q = R / L (L = length of shear wall)
R = w (tributary width) q = R / L (L = shear wall length
R1 = (150)16/2 = 1200 lbs q = 1200 / 8’ q = 150 plf
R2 = (150)(16+14)/2 = 2250 lbs q = 2250 / 12’ q = 188 plf
6
R3 = (150)14/2 = 1050 lbs q = 1050 / 8’ q = 131 plf
2 Rigid diaphragm
0 - 2 0 0
Concrete slabs and some steel decks are rigid; they transfer load in proportion to the
9 9
relative stiffness of each wall. Since rigid diaphragms experience only minor
1
deflections under load they impose equal drift on walls of equal length and stiffness.
erle
For unequal walls reactions are proportional to a resistance factor r.
c h ir = E I / h3 / ∑ (E I / h3)
h = wall height
GS
I = bL3 / 12 (moment of inertia of wall)
G
b = wall thickness
r i g h t L = wall length
For walls of equal height, thickness and material, the resistance factors are:
Copy r = L3 / Σ L3
L13 = L33 = 83 = 512
L23 = 123 = 1728
∑ L3 = 512 + 1728 + 512 Σ L3 = 2752
r1 = 512 / 2752 r1 = 0.186
r2 = 1728 / 2752 r2 = 0.628
r3 = 512 / 2752 r3 = 0.186
Check Σ r Σ r = 1.000
Total force F
F = 1000 plf x (16’+14’) / 1000 F = 30 k
Wall reactions
R1 = r1 F = 0.186 x 30 k R1 = 5.58 k
R2 = r2 F = 0.628 x 30 k R2 = 18.84 k
R3 = r3 F = 0.186 x 30 k R3 = 5.58 k
Check Σ R Σ R = 30,00 k

Example: Flexible diaphragm
IBC table 2306.4.1 excerpts Assume: plywood diaphragm, plywood shear walls on light wood framing
Allowable shear for wood panels with Douglas-Fir-Large or Southern Pine Dead load DL= 23 psf
Nail spacing at panel edge (inches) Seismic factor (adjusted for ASD) Cs = 0.15
Panel Panel Nail 6 4 3 2* Dead load per level
Nail size
grade thickness penetration Allowable shear (lbs / foot) W = 23 psf x 68’ x 150’ / 1000 W = 235 k
5/16 in 1 1/4 in 6d 200 300 390 510 Total DL (3 levels)
Structural I
sheathing
3/8 in 1 3/8 in 8d 230 360 460 610 ∑ W = 3 x 245 k ∑ W = 705 k

7/16 in 1 3/8 in 8d 255 395 505 670
Base shear
1 3/8 in 8d 280 430 550 730
15/32 in V = W Cs = 705 x 0.15 V = 106 k
1 1/2 in 10d 340 510 665 870
* Requires 3 x framing and staggered nailing Force distribution
Level Wx hx Wxhx wxhx/Σwihi Fx=V(wxhx/Σwihi) Vx =Σ Fx
2 235 k 27’ 6345 k’ 0.50 53 k 53 k
1
0
235 k
235 k
18’
9’
4230 k’
2115 k’
- 2 0 0
0.33
0.17
6 35 k
18 k
88 k
106 k
9
ΣwihI = 12,690 k’ V = 106 k
1 9
Area per level A = 68 (150)
9 A= 10200 ft2
erle
Shear per square foot v v =V/A v (psi)
c h i V0 = 106 k = 106000 lbs

V1 = 88 k = 88000 lbs
v0 = 106000/10200
v1 = 88000/10200
v0=10.4 psf
v1 = 8.6 psf
GS
V2 = 53 k = 53000 lbs v2 = 53000/10200 v2 = 5.2 psf
h t G Level 0 shear walls
rig
Wall A = 10.4 psf (15’)30’/12’ = 390 plf use 5/16, 6d @ 3” = 390 plf
C o p y Wall B = 10.4 psf (19’)30’/24’ =

Wall C = 10.4 psf (34’)15’/30’ =
Wall D = 10.4 psf (34’)30’/30’ =
Level 1 shear walls

247 plf
177 plf
354 plf
use 7/16, 8d @ 6” =
use 5/16, 6d @ 6” =
use 3/8, 8d @ 4” =
255 plf
200 plf
360 plf
Wall B = 8.6 psf (19’)30’/24’ = 204 plf use 3/8, 8d @ 6” = 230 plf
Wall C = 8.6 psf (34’)15’/30’ = 146 plf use 5/16, 6d @ 6” = 200 plf
Wall D = 8.6 psf (34’)30’/30’ = 292 plf use 5/16, 6d @ 4” = 300 plf
Level 2 shear walls

Wall B = 5.2 psf (19’)30’/24’ = 124 plf use 5/16, 6d @ 6” = 200 plf
Wall C = 5.2 psf (34’)15’/30’ = 89 plf use 5/16, 6d @ 6” = 200 plf
Wall D = 5.2 psf (34’)30’/30’ = 177 plf use 5/16, 6d @ 6” = 200 plf
Note: To simplify construction, fewer wall types could be selected

Example: Rigid diaphragm
Assume: concrete slab on CMU shear walls
Allowable masonry shear stress Fv = 85 psi
Seismic factor Cs = 0.17x1.5 Cs = 0.26
Note: increase Cs by 1.5 per IBC 2106.5.1 for ASD method
Dead Load
Wall lengths L = 12 (30’)+14 (12)+8 (24) L= 720’
Wall DL = (720’) 8’(7.625”/12”) 120 pcf / (68x150) DL = 43 psf
Floor/roof (12” slab) 150 psf
Miscellaneous 7 psf
Σ DL Σ DL = 200 psf
DL / level: W = 200 psf x 68’ x 150’/ 1000 W = 2,040 k
DL for 3 Levels: W = 3 x 2040 k W = 6,120 k
6
Base shear V = Cs W = 0.26 x 6120 V = 1,591 k
Force distribution
Level Wx hx Wxhx
0-2 0
wxhx/Σwihi
0 Fx= V(wxhx/Σwihi) Vx = Σ Fx
9
2 2,040 k 27’ 55,080 k’ 1591x0.50 796 k 796 k
1
1 9
2,040 k 18’ 36,720 k’ 1591x0.33 525 k 1,321 k
erle
0 2,040 k 9’ 18,360 k’ 1591x0.17 270 k 1,591 k
ΣwihI = 110,169 k’ V = 1,591 k
c h i
Relative wall stiffness:
GS
R = L3 / Σ L3
Wall B: r = 123 / [123+243] r = 0.11
i g h t G Wall C: r = 243 / [123+243]

Wall cross section areas:
r = 0.89
y r A walls = 12(30’)12”(7.625”) A = 32940 in2
Cop
B walls = 14(12’)12”(7.625”) B = 15372 in2
C walls = 8(24’)12”(7.625”) C = 17568 in2
Level 0 (V0 = 1591 k)
Wall A = (1591) 1000 / 32940 48 psi < 85
Wall B = (1591) 1000 (0.11) / 15372 11 psi < 85
Wall C = (1591) 1000 (0.89) / 17568 81 psi < 85
Level 1 (V1 = 1321 k)
Wall A = (1321) 1000 / 32940 40 psi < 85
Wall B = (1321) 1000 (0.11) / 15372 10 psi < 85
Wall C = (1321) 1000 (0.89) / 17568 67 psi < 85
Level 2 (V2 = 796 k)
Wall A = (796) 1000 / 32940 24 psi < 85
Wall B = (796) 1000 (0.11) / 15372 6 psi < 85
Wall C = (796) 1000 (0.89) / 17568 40 psi < 85

Seismic design issues
Eccentricity
Offset between center of mass and center of resistance causes eccentricity which causes
torsion under seismic load. The plans at left identify concentric and eccentric conditions:
1 X-direction concentric
Y-direction eccentric
2 X-direction eccentric
Y-direction eccentric
- 2 0 0 6
3 X-direction
1 9
concentric
9 0
erle
Y-direction concentric
i
GS c h Y-direction concentric
i g h t G
y r
Cop 5 X-direction concentric
Note:
Plan 5 provides greater resistance against torsion than plan 6 due to wider wall spacing
Plan 6 provides greater bending resistance because walls act together as core and thus
provide a greater moment of inertia

Hillside construction
To avoid expensive earthquake settlement repair ……
- 2 0 0 6
1 9 9 0
c h i erle
GS
………. adapt building to site rather than adapting site to building
i g h t G
y r
Cop

Critical wood-frame items
Item Requirements
A Shear wall anchor bolts Resist wall slippage
B Hold-down Resist shear wall overturning

A A B B
6" MAX C Shear wall nailing Attach panels to framing
6" MAX
12" MAX D Wall-to-wall hold-down Resist shear wall overturning
6
6" MAX
C C DD
0 - 2 0 0
E
1 9
Framing anchor clips
9
Transfer shear from floor to floor
E
c h i erle
E
h t G GS
y r i g F Beam connection Resist beam slippage
Cop
PLWD. GYP/ FI BER BD. G Shear wall width/height ratio Minimum 1 : 3.5 for stability
3.5 1.5 Wood panels 1:3.5 (Los Angeles, 1:2)
Gypsum board 1:2
F GG
1.0 1.0
F
H Joist blocking Transfers shear at panel edges
H H

Base Isolator
Left: Conventional structure
• Large total and inter-story drift
• Accelerations increase with height
• Potential permanent deformations
• Potential equipment damage
Right: Base isolators:

• Reduce floor accelerations and drift
• Reduce damage to structure and equipment
• Are not good for high-rise structures
- 2 0 0 6
Left:
9
Base isolator make-up
1
• Top and bottom steel plate
9 0
erle
• Rubber sheets
c h i • Steel sheets
• Central lead core
h t G GS Right: Separate building from ground to allow drift
y r i g
Cop
UCLA Kerckhoff Hall base isolator upgrade

Drawings, courtesy Widom Wein Cohen Architects, Santa Monica

System Selection
10
Structural design starts with the selection of a system and material; often informed by
similar past projects, even if not appropriate. For example, light wood structures are
common for residential building where hurricanes cause frequent destruction, though
heavy concrete or masonry would resist wind load much better. A rational method is
proposed with the objective to select more appropriate systems. However, since design
criteria may be conflicting in some cases, selection is both art and science, yet the
following criteria make the selection process more objective
• Capacity limit
Conc e pt ua l De sign • Code requirements
• Cost
• Load
• Location
• Resources
• Technology
- 2 0 0 6
• Synergy
9 9 0
Capacity limit is based on limits of systems and materials. For example, beams are
1
erle
economical for a given span range. To exceed that range would yield a bad ratio of dead
load to live load. A beam’s cross section increases with span, resulting in heavier dead
c h i load. Eventually, the beam’s dead load exceeds its capacity and it would break.
GS
Approaching that limit, the beam gets increasingly uneconomical because its dead weight
leaves little reserve capacity to carry live load. The span limit can be extended by
G
effective cross section shape. For example, steel beam cross sections are optimized in
r i g h t response to bending and shear stress, to allow greater spans.
y
Trusses have longer span capacity than beams, due to reduced self weight. They
Cop
replace the bulk of beams by top and bottom chords to resist global moments, and
vertical and diagonal web bars to transfer shear between compression and tension
chords. Compared to beams, the greater depth of trusses provides a greater lever arm
between compression and tension bars to resist global moments. Similarly, suspension
cables use the sag between support and mid-span as moment resisting lever arm. Since
cables have higher breaking strength and resist tension only, without buckling, they are
Introduction optimal for long spans; but the high cost of end fittings makes them expensive solutions
for short spans. These examples show, most systems have upper and lower span limits.
Conceptual design usually starts with approximate sizing principle elements of a structure
and possible alternatives, followed by thorough analysis during design development. Code requirements define structures by type of construction regarding materials and
Approximate methods are essential to quickly develop alternate designs. They are also systems; ranging form type I to type V for least and most restrictive, respectively, of the
useful to verify final designs and computer analysis. If based on good assumptions, Uniform Building Code (UBC) for example. Each type of construction has requirements
approximate methods can provide results of remarkable accuracy, usually within ten for fire resistance, maximum allowable floor area, building height, and occupancy group.
percent of precise results. The following conceptual design examples introduce Codes also have detailed requirements regarding seismic design; notable structures are
approximate methods, sometimes referred to as back-of-the-envelop design. They are categorized by ductility to absorb seismic energy and related height limits. Some code
not meant to replace accurate design but as precursor of accurate design and analysis. requirements are related to other criteria described in the respective section.
10-1 DESIGN METHODS Conceptual Design

Cost is often an overriding criterion in the selection of structures. In fact, cost is often conditions, namely continuity of demand to justify the high investments associated with
defined by some of the other selection criteria. However, costs also depend on market prefab concrete technology.
conditions and seasonable changes. The availability of material and products, as well as
Synergy, defined as a system that is greater than the sum of its parts is a powerful
economic conditions and labor strikes may greatly affect the cost of structures. For
concept to enrich architecture, regarding both pragmatic as well as philosophic
example, a labor strike in the steel industry may shift the advantage to a concrete
objectives. Pragmatic example are numerous: Wall system are appropriate for hotel and
structure, or the shortage of lumber, may give a cost advantage to light gauge steel
apartment projects which require spatial and sound separation; but moment frames
instead of light wood framing. Sometimes, several systems are evaluated, or schematic
provide better space planning flexibility as needed for office buildings. However, the core
designs are developed for them, in order to select the most cost effective alternative.
of office buildings, usually housing elevators, stairs, bathrooms, and mechanical ducts,
Load imposed on a structure is a major factor in selecting a system. For example, roofs
without the need for planning flexibility, often consists of shear walls ore braced frames,
in areas without snow must be designed only for a nominal load, yet roof load in
effective to reduce drift under lateral loads. Long-span systems provide column-free
mountain areas may be up to 20 times greater than the nominal load. Structures in
space required for unobstructed views in auditoriums and other assembly halls; but lower
earthquake prone areas should be lightweight and ductile, since seismic forces are
cost short span systems are used for warehouses and similar facilities where columns
basically governed by Newton’s law, force equals mass times acceleration (f=ma). In
are usually acceptable.
6
contrast, structures subject to wind load should be heavy and stiff to resist wind uplift and
minimize drift. Structures in areas of daily temperature variations should be designed for
thermal load as well, unless the structure is protected behind a thermal insulation skin
- 2 0 0
On a more detailed level, to incorporate mechanical systems within a long-span roof or
floor structure, a Vierendeel girder may be selected instead of a truss, since the
0
9
and subjected to constant indoor temperature only. rectangular panels of a Vierendeel better facilitate ducts to pass through than triangular
Location may effects structure selection by the type of soil, topography, and ground water
1 9
truss panels. A suspended cable roof may be selected for a sports arena if bleachers
erle
can be used to effectively resist the roof’s lateral thrust which is very substantial and may
level, natural hazards, such as fire, frost, or flood. Local soil conditions affect the
require costly foundations otherwise. Synergy is also a powerful concept regarding more
i
foundation and possibly the entire structure. Soft soil may require pile foundations; a mat
h
philosophical objectives, as demonstrated throughout history, from early post and beam
c
foundation may be chosen to balance the floating effect of high ground water. Locations
GS
structures; Roman arches, domes and vaults; Gothic cathedrals; to contemporary
with winter frost require deep foundations to prevent damage due to soil expansion in
suspension bridges or roofs. Columns can provide architectural expression as in post
frost (usually a depth of about one meter). Hillside locations may require caisson
G
and beam systems, or define and organize circulation, as in a Gothic cathedral. The
t
foundations to prevent sliding, but foundations are more common on flat sites. Locations
h
funicular surface of arches, domes and vaults can define a unique and spiritual space.
g
with fire hazards require non-combustible material. Raising the structure off the ground
y r
may be the answer to flooding.
i The buttresses to resist their lateral thrust provide the unique vocabulary of Gothic
Cop
cathedrals. Large retaining walls may use buttressing for rhythmic relieve, as in the great
Resources have a strong impact on the selection of structure materials. Availability of wall of Assisi, or lean backward to express increased stability as the wall of the Dalai
material was a deciding factor regarding the choice of material throughout history. The Lama palace in Tibet.
Viking build wood structures, a logical response to the vast forests of Scandinavia, yet
stone temples of Egypt and Greece reflect the availability of stone and scarcity of wood.
More recently, high-rise structures in the United States are usually steel structures, but
the scarcity of steel in some other countries makes concrete structures more common.
Technology available at an area also effects the selection of structures. For example,
light wood structures, known as platform framing, is most common for low rise residential
structures in the United States, where it is widely available and very well known; but in
Europe where this technology is less known, it is more expensive than more common
masonry structures. Similarly, in some areas concrete technology is more familiar and
available than steel technology. Concrete tends to be more common in areas of low
labor cost, because concrete form-wok is labor intensive. On the other hand, prefab
concrete technology is less dependent on low labor cost and more affected by market

Global moment and shear
Global moments help to analyze not only a beam but also truss, cable or arch. They all
resist global moments by a couple F times lever arm d:
M = F d; hence F = M / d
The force F is expressed as T (tension) and C (compression) for beam or truss and H
(horizontal reaction) for suspension cable or arch, forces are always defined by the global
moment and lever arm of resisting couple. For uniform load and simple support, the
maximum moment M and maximum shear V are computed as:
M = w L2 / 8
V =wL/2
w = uniform gravity load
L = span
- 2 0 0 6
9 9 0
For other load or support conditions use appropriate formulas
Beam
1
erle
Beams resist the global moment by a force couple, with lever arm of 2/3 the beam depth
c h i
d; resisted by top compression C and bottom tension T.
GS
Truss
G
Trusses resist the global moment by a force couple and truss depth d as lever arm; with
r i g h t compression C in top chord and tension T in bottom chord. Global shear is resisted by
vertical and / or diagonal web bars. Maximum moment at mid-span causes maximum
y
chord forces. Maximum support shear causes maximum web bar forces.
Cop Cable
Suspension cables resist the global moment by horizontal reaction with sag f as lever
arm. The horizontal reaction H, vertical reaction R, and maximum cable tension T form
an equilibrium vector triangle; hence the maximum cable tension is:
T = (H2 + R2)1/2
Arch
Arches resist the global moment like a cable, but in compression instead of tension:
C = (H2+R2)1/2
However, unlike cables, arches don’t adjust their form for changing loads; hence, they
assume bending under non-uniform load as product of funicular force and lever arm
between funicular line and arch form (bending stress is substituted by conservative axial
stress for approximate schematic design).

Radial pressure
Referring to diagram A, pressure per unit length acting in radial direction on a circular
ring yields a ring tension, defined as:
T=Rp
T = ring tension
R = radius of ring
p = uniform radial pressure per unit length
Units must be compatible, i.e., if p is force per foot, R must be in feet, if p is force per
meter, R must be in meters. Pressure p acting reversed toward the ring center would
reverse the ring force from tension to compression.
Proof
- 2 0 0 6
0
Referring to ring segment B:
1 9
T acts perpendicular to ring radius R
9
erle
p acts perpendicular to ring segment of unit length
c h i
Referring to ring segment B and vector triangle C:
GS
p and T in C represent equilibrium at o in B
h t G T / p = R / 1 (due to similar triangles), hence
rig
T=Rp
C o p y

Examples
Vierendeel Girder
Assume
Steel girders spaced 20’ L = 100’
Allowable stress (60% of Fy = 46 ksi tubing) Fa = 27.6 ksi
DL= 18 psf
LL = 12 psf (20 psf reduced to 60% for tributary area > 600 sq. ft.)
Σ = 30 psf
Uniform girder load
w = 30 psf x 20’ / 1000 w = 0.6 klf
Joint load
6
P= w e = 0.6 x 10’ P=6k
Vertical Reaction
R= w L / 2 = 0.6 x 100’ / 2
0 - 2 0 0 R = 30 k
9
END BAY CHORD
Chord shear
1 9
erle
Vc= (R – P/2) / 2 = (30 – 6/2) / 2 Vc = 13.5 k
i
Chord bending
GS c hMc = Vc e/2 = 13.5 x 10’x12” / 2

Moment of Inertia required
Mc = 810 k”
G
I= Mc c/ Fa = 810x5”/27.6 ksi I = 147 in4
r i g h t Use ST 10x10x5/16
WEB BAR (2nd web resists bending of 2 adjacent chords)
I = 183 > 147, ok
Copy 2nd bay chord shear

Vc= (R-1.5 P)/2 = (30–1.5x6)/2
2nd bay chord bending
Mc = Vc e/2 = 10.5x10’x12”/2
Vc = 10.5 k
Mc = 630 k”
Web bending
Mw = Mc end bay + Mc 2nd bay = 810 + 630 Mw = 1,440 k”
Moment of Inertia required
I = Mw c/Fa = 1,440x5/27.6 I = 261 in4
Use ST 10x10x1/2 web bar I =271 > 261, ok
MID-SPAN CHORD (small chord bending ignored)
Mid-span global bending
M= w L2/8= 0.6x1002/8 M = 750 k’
Mid-span chord force
P= M/d= 750/6 P= 125 k
Use ST 10 x 10 x5/16 297 > 125, ok

Wood Arch
Assume
Three-hinge glue-lam arches spaced 16’.
(Available glue-lam dimensions: ¾” lams; 31/8”, 51/8”, 6¾”, 8¾” and10 ¾” wide).
Allowable buckling stress (from case studies) Fc’ = 200 psi
LL = 12 psf (reduced to 60% of 20 psf for tributary area > 600 sq. ft.)
DL = 18 psf (estimate)
Σ = 30 psf
Uniform load
w = 30 psf x 16’ / 1000 = w = 0.48 klf
Global moment
M = w L2 / 8 = 0.48 x 1002 / 8 = M = 600 k’
Horizontal reaction
H = M / d = 600 / 20 =
Vertical reaction
- 2 0 0 6 H = 30 k
R = w L /2 = 0.48 x 100’ / 2 =
1 9
Arch compression (max.)
9 0 R = 24 k
rle
C = (H2 + R2)1/2 = (30 2 + 24 2 )1/2 C = 38 k
e
Cross section area required
S c h i
A = C / Fc’ = 38 / 0.2 ksi
Glue-lam depth (try 5 1/8” wide glue-lam)
A = 190 in2
G
t = A / width = 190 / 5.125 = 37; use 50 lams of ¾” t = 37.5”
G
Check slenderness ratio
rig h t L / t = 100’ x 12” / 37.5” = L/t = 32 ok
y
Note:
Co p Arch slenderness of L/t = 32 is ok (the 5 1/8” arch width is braced against buckling by the
roof diaphragm).
Wind bracing at end bays may consist of diagonal steel rods in combination with
compression struts. The lateral thrust of arches may be resisted by concrete piers that
may be tied together by grade beams to resist the lateral arch thrust.
Final design must consider non-uniform load (snow on half the arch) resulting in
combined axial and bending stress; the bending moment being axial force times lever
arm between funicular pressure line and arch center. The funicular line may be found
graphically.
Graphic method
• Draw a vector of the computed vertical reaction
• Draw equilibrium vectors parallel to arch support tangent
• Equilibrium vectors give arch force and horizontal reaction

Case studies
Skating Rink, Heerenveen, Holland
Architect: Van der Zee & Ybema
Engineer: Arie Krijegsman, ABT
Steel trusses
Allowable stress Fy = 36 ksi x 0,6 Fa = 21.6 ksi
Truss span L= 66m/0.3048 L = 217’
Truss spacing e =7.2m/.3048 e = 24’
Truss depth at mid span d = 5.8m/0.3048 d = 19’
DL = o.6 kPa (12.5 psf)
LL = 0.5 kPa (10.4 psf)
Σ = 1.1 kPa (22.9 psf)
Uniform load per truss
w = 24’ x 22.9 psf / 1000
2 0 0 6
Mid span point load (center truss, A transfers load of circular end units)
-
w = 0.55 klf
0
Tributary area of end units
A = π r2/3 = π(217’/2)2/3
1 9
Point load per truss
9 A = 12,278 sq. ft.
erle
P = 12,278 x 22.9 psf / 1000 / 16 trusses P = 18 k
c h i
Global moment
M = PL/4+wL2/8 = 18x217/4 + 0.55 x 2172/8 M = 4,214 k’
GS
Chord bar force
G
C = T = M/d = 4,214 / 19 C =T = 222 k
r i g h t Bottom tension chord

Try wide flange section W8x35
y
Try wide flange
Cop
Allowable force P from AISC table (use L = 0’ for tension, no buckling)
Pall = 222 222 = 222, ok
Top chord un-braced length L = 217’/12 L = 18’
Top chord bending (negative support bending
M = w L2/12 = 0.55 x 182 / 12 M = 15 k’
Try W12x50
A = 14.7in2, Ix= 394 in4, rx= 5.17” (y-axis is braced by roof deck)
Bending stress
fb = M c / I = 15k’x12”x 6”/394 fb = 2.74 ksi
Axial stress fa = C / A = 222 k /14.7 in2 fa = 15.1 ksi
Slenderness KL/rx = 1x18’x12” / 5.17” kL/r = 42
Allowable buckling stress (from AISC table) Fa =19 ksi
Check combined stress fa/Fa + fb/Fb <=1
fa/Fa + fb/Fb = 15.1/19 + 2.74 / 21.6 = 0.92 0.92 < 1, ok
Use W12x50

Exhibit Hall 26 Hanover
Architect: Thomas Herzog
Engineer: Schlaich Bergermann
Given
Steel suspender bands 30x400 mm (1.2x16”), spaced 5.5 m (18’)
LL = 0.5 kN/m2 (10 psf)
DL = 1.2 kN/m2 (25 psf)
Σ= 1.7 kN/m2 (35 psf)
Uniform load
w = 1.7 kN/m2 x 5.5 m = w = 9.35 kN/m
Global moment
6
M = w L2 / 8 = 9.35 x 642 / 8 M = 4787 kN-m
Horizontal reaction
H = M / f = 4787 / 7
0 - 2 0 0 H = 684 kN
9
Vertical reaction R (max.)
1 9
erle
Reactions are unequal; use R/H ratio (similar triangles) to compute max. R
R/H = (2f+h/2) / (L/2), hence
c h i
R = H (2f+h/2) / (L/2) = 684 (2x7+13/2) / (64/2) 438 kN
GS
Suspender tension (max.)
T = (H2 + R2)1/2 = (684 2 + 438 2 )1/2 T = 812 kN
h t G Suspender stress (A = 30x400 mm)
rig
f = T / A = 1000x812 / (30x400) f = 67.7 MPa
C o p y US unit equivalent
67.7 kPa x 0.145 f = 9.8 ksi
9.8 < 22 ksi, ok
Graphic method
• Draw a vector of the total vertical load
• Equilibrium vectors parallel to support tangents give cable forces
• Equilibrium vectors at supports give H and R reactions.
Note: The unequal support height is a structural disadvantage since the horizontal
reactions of adjacent bays don’t balance, but it provides lighting and ventilation, a major
objective for sustainability. The roof consists of prefab wood panels, filled with gravel to
resist wind uplift. Curtain wall mullions at the roof edge are prestressed between roof
and footing to prevent buckling under roof deflection. In width direction the roof is slightly
convex for drainage; which also gives the interior roof line a pleasing spatial form.

Oakland Coliseum
Architect/Engineer: Skidmore Owings and Merrill
Assume
Allowable cable stress (210 ksi breaking strength / 3) Fa = 70 ksi
Radial suspension cables, spaced 13’ along outer compression ring
LL= 12 psf (60% of 20 psf for tributary area > 600 sq. ft.)
DL= 46 psf (estimate)
Σ= 58 psf
Uniform load
w = 58 psf x 13’ / 1000 w = 0 to 0.75 klf
Global moment [cubic parabola with origin at mid-span]
Mx = w L2/24 (1 – 8 X3 / L3 ) for max. M at mid-span, X=0, hence
M = w L2 / 24 = 0.75 x 4202 / 24
Horizontal reaction
- 2 0 0 6 M = 5,513 k
0
H = M / f = 5,513 / 30 H = 184 k
Vertical reaction
1 9 9
rle
R = w L / 2 = (0.75/2) x 420’ / 2 R = 79 k
e
Cable tension (max.)
Sc h i
T = (H2 + R2)1/2 = (184 2 + 79 2 )1/2
Metallic cross section required
T = 200 k
h t G G Am = T / Fa = 200 / 70 ksi
Gross cross section (70% metallic)
Am= 2.86 in2
p y ri g Ag = Am / 0.70 = 2.86 / 0,70

Cable size
Ag = 4.09 in2
Co
ø = 2(Am / π)1/2 = 2(4.09 / 3.14)1/2 = 2.28” use ø 2 3/8”
Steel tension ring (inner ring radius r= 15’, cable spacing = 0.94’)
T= H r / 0.94 = 184 x 15 / 0.94 T = 2,972 k
Cross-section area (assume high-strength steel Fa = 30 ksi)
A= T/Fa = 2,936/30= 98 in2
Try W24x335, A= 98.2 in2 > 98 use W24x335
Concrete compression ring (r = 210’, e = 13’)
C = H r / e = 184 x 210 / 13 = C = 2,972 k
Cross-section area (assume allowable buckling stress Fc’= 1.2 ksi)
A = C / Fc’ = 2,972 / 1.2 = 2,477 in2 ~72”x34” Use 6x3’

Olympic Stadium Munich
Architect: Guenter Behnisch
Engineer: Leonhardt und Andrae
Assume
Allowable cable stress (210 ksi breaking strength / 3) Fa = 70 ksi
DL = 5psf 5 psf
LL = 20 psf wind uplift 21 psf
Σ= 25 psf net uplift 16 psf
Uniform load (cable spacing 75 cm = 2.5’)
Gravity w = 25 psf x 2.5’ w = 62.5 plf
Wind uplift 16 psf x 2.5’ p = 40 plf
Global moment
6
M = w L2 / 8 = 62.5 x 1972 / 8 M = 303,195 #
Horizontal reaction
H = M / f = 303,195 #” / 39’
0 - 2 0 0 H = 7,774 #
9
Vertical reaction
1 9
erle
R = w L / 2 = 62.5 x 197’ / 2 R = 6,156 #
Gravity tension (add 10% residual prestress)
c h i
T = 1.1 (H2 + R2)1/2 = 1.1 (7,774 2 + 6,156 2 )1/2 Gravity T = 10,908 #
GS
Wind uplift tension (add 10% residual prestress)
Wind suction is normal to surface, hence T = p r [r = curvature radius)
i g h t G T = 1.1 p r = 1.1 x 40 x 226’ Wind T = 9,944 #
r
10,907 > 9,944 gravity governs
Copy Metallic cross section area (assume twin ½” net cables, 70% metallic)
Am = 0.7 x 2 π r2 = 2 x 3.14 (0.5/2)2
Net cable stress
f = T/ Am = 10,908 / (0.28 x 1000)
Am= 0.28 in2
f = 39 ksi < 70 ok
Plexiglas joint / support Net cable joint

Imos Factory Newport UK
Architect: Richard Rogers
Engineer: Anthony Hunt
This microprocessor factory is a cable stayed structure. Stay rods, instead of stay
cables, are suspended from trussed steel pylons along a central circulation spine. The
rods support prismatic roof trusses at third points to reduce truss depth and weight for a
column-free floor area of maximum flexibility. Truss joists, spaced 6m (20‘), support the
roof deck. Mechanical equipment, located over the central spine, also allows for optimal
flexibility, as required for this facility.
Assume
Allowable rod stress (high strength steel) Fa = 207,000 kPa (30 ksi)
DL = 0.7 kPa (14.7 psf)
LL =
Σ =
0.5 kPa
1.2 kPa
(10.4 psf)
(25.1 psf)
- 2 0 0 6
Uniform load
1 9
w = 1.2 kPa x 13.2m
9 0 w = 15.8 kN/m
erle
Tributary load per rod P
i
P = w L’ = 15.8 kN/m x 12m P = 190 kN
GS c hRod tensions
TA-B (from vector triangle) TA-B = 370 kN
i g h t G TA-C (from vector triangle)

Rods A-B cross section (2 rods)
TA-C = 248 kN
y r A = TA-B / (2 Fa) = 370 / (2 x 207,000) = 0.000894 m2 A = 894 mm2
Cop
Rod diameter
∅ = 2 (A/π)1/2 = 2 (894/3.14)1/2 2 ∅ 34 mm
Convert to inches
34 mm/25.4 2 ∅ 1.33”
Rods A-C cross section (4 rods)
A = TA-C / (4 Fa) = 246 / (4 x 207,000) = 0.000297 m2 A = 297 mm2
Rod diameter ∅
∅ = 2 (A/π)1/2 = 2 (297/3.14)1/2 4 ∅ 20 mm
Covert to inches
20 mm/25.4 4 ∅ 0.79”
Note: design the truss for added compression, from vector triangles
Between B-C add 317 kN B-C = 317 kN
Between C-D add 317 + 159 kN C-D = 476 kN

Sports Center UC Berkeley
Architect: G G Schierle
Engineer: T Y Lin
Assume:
Cable truss with vertical compression struts
Span L= 120’
Depth d = 10’
Spacing e = 20’
Allowable cable stress (210 ksi / 3) Fa= 70 ksi
Prestress 60% of Fa (50% + temperature change reserve)
DL = 18 psf
LL = 12 psf (60% of 20 psf for tributary area > 600 sq. ft.)
Σ = 30 psf
Uniform load per truss
- 2 0 0 6
w = 30 psf x 20’/1000
1 9
Global moment (fixed support)
9 0 w = 0.6 klf
erle
M = wL2/12 = 0.6 x 1202/12 M = 720 k’
c h i
Chord force (assume 10% residual prestress)
GS
T = 1.1 M/d = 1.1 x 720/10’ T = 79 k
Chord cross section area (70% metallic)
h t G A= T/(0.7 Fa) = 79/(0.7 x 70ksi) A =1.61 in2
rig
Chord cable size
y
φ = 2(A/π)1/2 =2(1.61/π)1/2 =1.43” Use φ1.5”
C o p Vertical reaction (without guy cable force)

R = wL/2 = 0.6 x 120’/2
Diagonal cable force (assume 10% residual prestress)
T = 1.1 T’ = 1.1 x 65 k (from vector triangle)
R = 36 k
T = 72 k
Diagonal cable cross section (twin cables, 70% metallic)
A = T’/(2 x 70 x 0.7) = 72/(2 x 70 x 0.7) A = 0,73 in2
Diagonal cable size
φ = 2(A/π)1/2 = 2(0.73/π)1/2 = 0.96 in Use φ1 in

Portal method
The Portal Method for rough moment frame design is based on these assumptions:
• Lateral forces resisted by frame action
• Inflection points at mid-height of columns
• Inflection points at mid-span of beams
• Column shear is based on tributary area
• Overturn is resisted by exterior columns only
1 Single moment frame (portal)
2 Multistory moment frame
3 Column shear is total shear V distributed proportional to tributary area:
Va = (V/B) L1 / 2
Vb = (V/B) (L1+L2) / 2
Vc = (V/B) (L2+L3) / 2
Vd = (V/B) L3 / 2
- 2 0 0 6
4
9 9 0
Column moment = column shear x height to inflection point
1
erle
Ma = Va h / 2
Mb = Vb h / 2
c h i Mc = Vc h / 2
GS
Md = Vd h / 2
5 Exterior columns resist most overturn; the portal method assumes they resist all
i g h t G 6 Overturn moments per level are the sum of forces above the level times lever arm
r
of each force to the column inflection point at the respective level:
Copy M2 = F2 h2 / 2 (level 2)
M1 = F2 (h2+h1 / 2) + F1 h1 / 2 (level 1)
Column axial force = overturn moment divided by width B
N=M/B
Column axial force per level:
N2 = M2 / B (level 2)
N1 = M1 / B (level 1)
7 Beam shear = column axial force below beam minus column axial force above beam
Level 1 beam shear:
V = N1 - N2
Roof beam:
V = N2 - 0 = N2
Beam bending = beam shear times distance to inflection point at beam center
M=VL/2
Beam axial force is negligible and assumed 0

Example: 2-story building 1st floor
Assume: Base shear
L1 = 30’ V = F1+F2 = 8+12 V = 20 k
L2 = 20’ Column shear
B = 30+20+30 = 80’ Va = (L1/2) (V/B) = 15’x20/80 Va = 3.75 k
h = h1 = h2 = h = 14’ Vb = (L1+L2)/2 (V/B) = (20+30)/2 (20/80) Vb = 6.25 k
F1 = 8 k Column bending
F2 = 12 k Ma = Va h/2 = 3.75 x 14/2 Ma = 26 k’
Mb = Vb h/2 = 6.25 x 14/2 Mb = 44 k’
Overturn moments
M1 = F2 (h2+h1/2)+F1 h1/2 = 12 (14+7)+8x7 M1 = 308 k
M2 = F2 h2/2 = 12x7 M2 = 84 k’
Column axial load 1st floor
6
N1 = M1/B = 308/80 N1 = 3.9 k
Column axial load 2nd floor
N2 = M2/B = 84/80
0-2 0 0 N2 = 1.1 k
9
Beam shear
1 9
V1 = N1–N2 = 3.9-1.1 V1 = 2.8 k
erle
Beam bending
M1 = V1 L1/2 = 2.8x30/2 M1 = 42 k’
c h i
M2 = V1 L2/2 = 2.8x20/2 M2 = 28 k’
GS
2nd floor
2nd floor shear
G
V = F2 V = 12 k
rig h t Column a shear

Va = (L1/2) (V/B)=15’x12/80 Va = 2.25 k
C o p y Column b shear
Vb = (L1+L2)/2 (V/B) = (30+20)/2x12/80
Column a bending
Ma = Va h/2 = 2.25 x 14/2
Vb = 3.75 k
Ma = 15.75 k’
Column b bending
Mb = Vb h/2 = 3.75 x 14/2 Mb = 26.25 k’
Overturn moment
M2 = F2xh/2 = 12x14/2 M2 = 84 k’
Column axial load
N2 = M2/B = 84/80 N2 = 1.0 k
Beam shear
V2 = N2 V2 = 1.0 k
Beam 1 bending
M1 = V2 L1/2=1.0x30/2 M1 = 15 k’
Beam 2 bending
M2 = V2 L2/2=1.0x20/2 M2 = 10 k’

Moment frame
Eight story steel moment frame, high strength steel, Fy = 50 ksi Fa = Fb = 30 ksi
Live load 50 psf, load reduction R in percent per IBC
R = 0.08 (A-150)
A = tributary area
Max. reduction: 40% for members supporting a single level, 60% for other members
Gravity load Beam (psf) Column (psf)
Framing 10 10
Concrete slab 37 37
Partitions 20 20
Floor / ceiling 3 3
DL 70 70
LL 50 x 0.6 = 30 50 X 0.4 = 20
Total DL + LL
Average wind pressure
- 2 0 0 6 100 90
P = 33 psf
0
Design ground floor and 4th floor
1 9
Uniform beam load (shaded tributary are)
W = 100 psf x 30’/1000
9 w = 3 klf
erle
Uniform column load (distributed on beam)
i
w = 90 psf x 30’/1000 w = 2.7 klf
GS c hBase shear
V = 33 psf x 30’ x 7.5 x 12’/1000
Level 4 shear
V = 89 k
i g h t G V = 33 psf x 30’ x 3.5 x 12’/1000

Overturn moments
V = 42 k
y r Ground floor M0 = 33 psf x 30’ x (7.5x12)2 /2 / 1000 M0 = 4,010 k’
Cop
First floor M1 = 33 psf x 30’(6.5 x12)2 /2 / 1000 M1 =3,012 k’
Fourth floor M4 = 33 psf x 30’(3.5 x12)2 /2 / 1000 M4 = 873 k’
Beam design
Column a & d axial load
N0 = M0 / B = 4,010 / 90 N0 = 45 k
N1 = M1 / B = 3,012 / 90 N1 = 34 k
Beam design
Beam shear
V = N0 – N1 = 45-34 V = 11 k
Beam bending
M lateral = V L/2 = 11x30/2 M lateral = 165 k’
M gravity = wL2/12 = 3x302/12 M gravity = 225 k’
ΣM = 165 + 225 Σ M = 390 k’
Required Sx = M / Fb = 12”x390 k’/ 30 ksi Sx = 156 in3
Use W18x86 Sx = 166 > 156
Note: W18 beam has optimal ratio L/d = 20

Ground floor
Column shear
Column Vc = L tributary V / B Vc
a&d 15x89/90 14.8 k
b&c 30x89/90 29.7 k
Column bending
Column M lateral = Vc h/2 M gravity = wL2/24 ΣM
a&d 14.8 x12/2 = 89 k’ 2.7x302/24 = 101 k’ 190 k’
b&c 29.7x12/2 = 178 k’ 0 178 k’
Column axial force (n = # of stories)
Column P lateral = Mo/B Pgravity = n w L tributary ΣP
a&d 4,010 / 90 = 45 k 8x2.7x15 = 324 k 369 k
b&c 0 8x2.7x30 = 648 k 648 k
6
Column axial force + bending (ΣP = P+M Bx, estimate Bx than verify)
Column
a&d
P
365 k
0 - 2 0 0 M Bx (convert M to k”)
12”x190 k’x0.185 = 422 k
ΣP
787 k
9
b&c 648 k 12”x178 k’x0.185 = 395 k 1043 k
1 9 Design column (assume KL = 1.2 x12 = 14’)
erle
Column Use Check P allowable vs. P Check Bx estimate vs. Bx
a&d W14x109 803 > 785, OK 0.185 = 0.185, OK
c h i
b&c W14x145 1090 > 1043, OK 0.185 > 0.184, OK
GS
4th floor
G
Column shear
t
Column Vc = L tributary V / B Vc
y r i g h a&d
b&c
15x42/90
30x42/90
7k
14 k
Cop
Column bending
Column M lateral = Vc h/2 M gravity = wL2/24 ΣM
a&d 7 x12/2 = 42 k’ 2.7x302/24 = 101 k’ 143 k’
b&c 14x12/2 = 84 k’ 0 84 k’
Column axial force (n = # of stories)
Column P lateral = Mo/B P gravity = nwL tributary ΣP
a&d 873/90 = 10 k 4x2.7x15 = 162 k 172 k
b&c 0 4x2.7x30 = 324 k 324 k
Column axial force + bending (ΣP = P+M Bx, estimate Bx than verify)
Column P M Bx (convert M to k”) ΣP
a&d 172 k 12”x143 k’x0.196 = 336 k 508 k
b&c 324 k 12”x84 k’x0.196 = 198 k 522 k
Design column (assume KL = 1.2 x12 = 14’)
Column Use Check P allowable vs. P Check Bx estimate vs. Bx
a&d W14x82 515 > 508, OK 0.196 = 0.196, OK
b&c W14x90 664 > 522, OK 0.196 > 0.185, OK

Braced frame
Eight story braced frame: high strength steel, Fy = 50 ksi Fa = Fb = 30 ksi
Loads
Gravity load Column Beam
DL 70 psf 70 psf
LL 50x0.4 = 20 psf 50x0.6 = 30 psf
Total DL+LL 90 psf 100 psf
Average wind pressure P = 30 psf
51’ Beam load
w = 100 psf x 30’/1000 w = 3 klf
Column load (per foot on beam)
w = 90 psf x 30’/1000 w = 2.7 klf
Base shear
0 0 6
0-2
V = 30 psf x 45 x 90’/1000 V = 122 k
Overturn moment
9 9
Lever arm (to floor level for braced frames)
1
rle
L = (8x12’-6’)/2+6’ L = 51’
M0 = V L = 122 k x51’ M0 = 6,222 k’
Sc h i e
Column
Column and brace axial forces
Plateral = M0/30’ Pgravity = n w Atributary ΣP
h t G G
Gravity tributary width = 30’
a&d
b&c
P=0
P = 6,222 k’/30’ = 207 k
P = 8x2.7x15 = 324 k
P = 8x2.7x30 = 648 k
324 k
855 k
i g
Brace See vectors (tension & compression, design for compression) 78 k
o p y r Column and brace design (K = 1 for pin joints)
C Wind tributary width = 45’

Column
a&d
b&c
Brace
Force P
324 k
855 k
KL length (K=1)
12’
12’
78 k L = (122+152)1/2 = 20’
W14x61
W14x120
TS6x6x5/16
Use Pallowable vs. P
410 > 324
919 > 855
93 > 78
Beam *aldesign
Bending moment
M = wL2/8 = 3 klf x 302/8 M = 338 k’
Section Modulus
S = M/Fb = 12” x 338 k’ / 30 ksi S = 135 in3
Use W18x76 146 > 135
P = 78 k P = 78 K Deflection Δ = (5/384) W L3/ (E I)
Δ = (5/384) (3x30) (30x12)3/ (30,000x1330) Δ = 1.37”
Allowable Δ = L/240 = (30’x12”)/240 = 1.5 1.5 > 1.37, ok
Note: ignore brace beam support of inner beam since lateral load may act in addition to gravity load
V = 122 k

4th floor design
4th floor shear
V = 33 psf x 45 x 3.5x12 / 1000 V = 62 k
Overturn moment
27’ Lever arm (to floor level for braced frames)
L = 3.5x12’/2+ 6’ L = 27’
M4 = V L = 62 k x27’ M4 = 1,674 k’
Column and brace axial forces
Column Plateral = M0/30’ Pgravity = n w Atributary ΣP
a&d P=0 P = 4x2.7x15 = 162 k 162 k
b&c P = 1,674 k’/30’ = 56 k P = 4x2.7x30 = 324 k 380 k
Brace See vectors (tension & compression, design for compression) 40 k
Column Force P KL length (K=1)
- 0 6
Column and brace design (K = 1 for pin joints)
2 0 Use Pallowable vs. P
0
a&d 162 k 12’ W8x31 189 > 162
b&c
Brace
1 9
380 k
9 12’
40 k L = (122+152)1/2 = 20’
W14x61
TS6x6x3/16
410 > 380
60 > 40
erle
Compare material
c h i
The amount of steel required per square foot (m2) is used to compare framing systems.
GS
The steel at mid height provides a quick average weight to compare, assuming all bays
of approximately the same size provides the following comparative results.
i g t G
Tributary width for gravity = 30’
h
Gravity tributary width = 30’ Member
Moment frame
Weight / ft Length each Total length Weight
y r 8 columns W14x82 82 plf 12’ 96’ 7,872 #
Cop
8 columns W14x90 90 plf 12’ 96’ 8,640 #
24 beams W18x86 86 plf 30’ 720’ 61,920 #
18 joists W18x35 35 plf 30’ 540’ 18,900 #
Total 97,332 #
Total per square foot 97,332 / (90x90) 12 psf
Tributary width for wind = 45’
Wind tributary width = 45’
Braced frame
Member Weight / ft Length each Total length Weight
8 columns W8x31 31 plf 12’ 96’ 2,976 #
8 columns W14x61 61 plf 12’ 96’ 5,856 #
24 beams W18x76 76 plf 30’ 720’ 54,720 #
18 joists W18x35 35 plf 30’ 540’ 18,900 #
P = 40 k P = 40 K 4 braces TS6x6x3/16 14.53 plf 20’ 80’ 1,162 #
Total 83,614 #
Total per square foot 83,614 / (90x90) 10.3 psf
V = 62 k

Test models
Static models are useful to test structures for strength, stiffness, and stability. They may
have axial resistance (truss), bending resistance (beam), or both axial and bending
resistance (moment frame). Static models have three scales: geometric scale, force
scale, and strain scale. The geometric scale relates model dimensions to original
dimensions, such as 1:100. The force scale relates model forces to the original structure.
For a force scale of 1:100, one pound in the model implies 100 pounds in the original
structure. The force scale should be chosen to keep model forces reasonable (usually
less than 50 pounds). The strain scale relates model strain (deformation) to strain in the
original structure. A strain scale of 1:1 implies model strain relates to original strain in the
geometric scale; given a geometric scale of 1:10 a model strain of 1 inch implies 10 inch
original strain. For structures with small deformations may a strain scale of 5:1, for
example, helps to visualize strain. However, structures with large strain like membranes
Geometric Scale:
- 2 0 6
require a strain scale of 1:1 to avoid errors (see 2). Scales are defined as:
0
SG = Lm/Lo = model dimension / original dimension
Force Scale:
1 9 9 0
SF = Pm/Po = model force / original force
SS = εm/εm = model strain / original strain
erle
Strain Scale:
The derivation for axial and bending resistance models assumes:
c h i A = Cross-section area
GS
E = Modulus of elasticity
I = Moment of inertia
i g h t G k = Constant of integration for deflection; for cantilever beams with

point load Δ = kPL3 / (EI) where k = 1/3
y r m = Subscript for model
Cop
o = Subscript for original structure
Axial resistance
Unit Strain ε = ΔL/L ΔL = P L / (AE) hence
Force P = A E ΔL/L = A E ε hence
Force Scale = SF = Pm/Po = AmEm / (AoEo) εm/εo since εm/εo = SS
SF = AmEm/(AoEo) SS
SF = AmEm/(AoEo) if SS = 1
SF = Am/Ao = SG2 if = Em=Eo
1 Axial strain ΔL = P L/(AE); unit strain ε = ΔL/L
2 Structures with large deformations, such as membranes, yield errors if the strain
scale SS is not 1:1; as demonstrated in the force polygon
3 Structures like trusses, with small deformations, may require a strain scale SS >1 to
better visualize deformations.

Bending resistance
Unit Strain ε = Δ/L Δ = kPL3 / (EI) hence
Force P = EIΔ / (kL3) = EI / (kL2) Δ/L hence
Force Scale SF = Pm/Po = [EmIm / (EoIo)] ko/km Lo2/Lm2 εm/εo
Since the model and original have the same load and support conditions the constants of
integration km = ko, hence the term ko/km may be eliminated. The term Lo2/Lm2 = 1/SG2 =
1/geometric scale squared, and εm/εo = SS = strain scale. Therefore the force scale is:
SF = EmIm/(EoIo) 1/SG2 SS
SF = EmIm/(EoIo) 1/SG2 If SS =1
SF = Im/Io 1/SG2 If Em = Eo, or simplified
SF = SG2
- 2 0 0 6
assuming all model dimensions,
0
including details, relate to the original in the geometric scale
1 9 9
In the simplest form the force scale is equal to the geometric scale squared for both axial
erle
and bending resistant models. Thus a model with a geometric scale of 1:100 has a force
scale of 1:10,000 if it is made of the same material or modulus of elasticity as the original
c h i
structure.
GS
Combined axial and bending resistance
G
Models with both axial and bending resistance, such as moment frames, should be of the
r i g h t same material or elastic modulus as the original in order to avoid errors. Referring to
diagrams 3, if, for example, the elastic modulus of a model is half as much as in the
y
original structure and the cross-section area is doubled to compensate for it, then the
Cop
moment of inertia is four times greater, assuming area increase is perpendicular to the
bending axis. For small adjustments this can be avoided by increasing the area parallel
to the bending axis. Large differences in stiffness, such as wood simulating steel, with an
elastic modulus about 20 times greater, are not possible. In such a case the strain scale
could be 20:1 to amplify deflection rather than adjusting the cross-section area.
1 Model strain ε = Δ/L must be equal to the original strain.
Δ = k PL3/EI where k = 1/3
2 Original strain ε = Δ/L
Δ = k PL3/EI where k=1/3
Since k is the same in the model as in the original, for equal load and support
conditions, it may be eliminated from the force scale equation
3 Correlation between cross-section area A and moment of inertia I
demonstrates incompatibility between A and I since they increase at different rates,
unless the increase is only in width direction

Test stand
• Light gauge steel frame 3’x5’
• Frame to support test models
• Adjustable platform for loads below the test model
(crank mechanism lowers platform to apply load)
• Blocking deice holds load platform at any position
• Support frame for measure gauges above test model
Test procedure
• Position model with open base to allow loads below
• Suspend load cups filled with lead or sand from model
(support loads on load platform before loading)
• Attach measure gauges above model
• Lower load platform with crank to apply load
• Measure deformations and stress
• Apply alternate loads (half load, etc.)
- 2 0 0 6
Note
1 9 9 0
• Record deformations and stresses for all load conditions
erle
• Apply loads briefly to avoid creep deformation
i
• Apply loads gradually to avoid rupture
GS c h• Test all load conditions that may cause critical deformation or stress
• Adjust design if deformation or stress exceeds acceptable limits
i g h t G
y r
Cop

For most material the E modulus may be defined, applying load on a cantilever and
computing E as:
E = PL3/(IΔ)
For fabric the E modulus is defined, applying axial load on a 5 inch fabric strip:
E = PL/(AΔ)
E = Elastic modulus in lbs/in2 (lbs/linear in for fabric)
Δ = deformation
A = width of fabric strip
Tree structure wood design model P = point load
L = length of cantilever or fabric strip
I = moment of inertia
Note:
L should be as long as possible for accuracy
6
Use average E of several tests
0 - 2 0 0
1 9 9
erle
Master thesis by Madhu Thangavelu
i
Wood cantilever roof Wood grid shell
h
MALEO: Modular Assembly in Low Earth Orbit, to avoid assembly of lunar station by
GS c costly remote control robotics. Light-weight cable truss of stable triangular configuration
supports three fuel tank modules for habitation, research, power and control.
i g h t G
y r
Cop Arch / grid shell Folded truss
Prismatic truss Cable stayed roof MALEO: Model

Computer aided design
Advance in computer technology made structural design and analysis widely available.
The theory and algorithm of structural design programs is beyond the scope of this book.
However, a brief introduction clarifies their potential and use.
Structure programs generate and solve a stiffness matrix of the structure. Based on the
degree of freedom of joints, the output provides stress and strain. A two-dimensional
truss with pin joints has two degrees of freedom and thus two unknowns per joint, X and
Y-displacement. A three-d truss has three unknowns per joint. Two-D frames have four
Moment frame (with joint numbers and member names) unknowns, X, Y-displacement and X, Y-rotation, but three-D frames have six unknowns
per joint, X, Y, Z-displacement and X, Y, Z-rotation.
The structure input is defined by joints, members connecting the joints and loads. Joints
of three-d structures are defined by X, Y, Z-coordinates, joint type (pin or moment joint),
2 0 0 6
and degree of freedom, regarding X, Y Z-displacement and X, Y, Z-rotation (joints
attached to the ground are fixed with pin or moment joints). Members are defined by
-
0
properties, cross section area, moment of inertia, and modulus of elasticity. Some
9 9
members may have end release at one or both ends, to allow pin joints of braces to
1
connect to moment joints of beam to column, for example. End releases are simulated by
erle
a dummy pin adjacent to the moment joint. The geometry of a structure may be defined
i
in the analysis program or imported as DFX file from a CAD program. Loads are defined
GS
Load diagram (uniform beam load, lateral point load)
c has distributed or point load. Gravity load is usually assigned as uniform beam load, yet
lateral wind or seismic loads are usually assigned as point loads at each level.
i g h t G Output includes force, stress, and deformation for members, joint displacement and
rotation, as well as support reactions. Output may be in tables and / or graphic display.
r
Graphic display provides better intuitive understanding and is more convenient to use.
Cop y Some programs simulate non-linear material behavior and / or non-linear geometric
behavior. For example, non-linear material may include plastic design of steel with non-
linear stress/strain relation in the plastic range. Non-linear geometric analysis is for
structures with large displacements, such as cable or membrane structures. Non-linear
analysis usually involves an iterative algorithm that converges after several iterations to a
desired level of accuracy. Some programs include a prestress element to provide form-
finding for membranes structures. Some programs provide dynamic analysis, sometimes
referred too as 4-D analysis. Programs with advanced features provide greater versatility
and accuracy, but they are usually more complex to use.
Multiframe-4D used for the demonstrations features 2-d and 3-d static and 4-d dynamic
analysis. For static analysis Multiframe is very user friendly, intuitive, and thus good for
architecture students. The 4-d dynamic feature is beyond the scope of this book. The
examples presented demonstrate 2-d and 3-d design/analysis. A very convenient feature
are tables of steel sections with pre-defined properties for US sections and for several
Hexagonal grid shell dome other countries. The program features US and SI units.

Belt truss effect
CAD-analysis provides efficient means to compare framing systems. For convenience
the following example was done with constant W18 beams and W14 columns, 30’ beam
spans and 12’ story heights. The results, comparing the effect of belt and top trusses on
a moment frame and a braced frame are very reveling:
20-story moment frame

Gravity load w = 3 klf
Wind load P = 10 k / level
6
Frame: Drift
Frame only
Top truss
Belt truss
15.1”
0 -
14.9”
14.2”
2 0 0
9
Top and belt truss
1 9
14.0”
c h i erle
h t G GS 20-story braced frame
Gravity load w = 3 klf
y r i g Wind load P = 10 k / level
Cop
Frame Drift
Frame only 17.6”
Top truss 11.4”
Belt truss 11.1”
Top and belt truss 8.6”
Note:
Belt and top trusses are much more effective to reduce drift at the braced frame than at
the moment frame. The combined belt and top trusses reduce drift:
• 7 % at moment frame
• 49 % at braced frame
Interpreting the results clarifies the stark difference and fosters intuitive understanding of
different deformation modes of moment and braced frames.

PART I V 11
H ORI Z ON T AL SY ST EM S H ORI Z ON T AL SY ST EM S
Part III presents structure systems, divided into two categories: horizontal, and
Be nding Re sist a nt
vertical/lateral. Horizontal systems include floor- and roof framing systems that support
gravity dead- and live load and transfer it to vertical supports, such as walls and columns.
- 2 0 0 6
0
As the name implies, vertical/lateral systems include walls, columns and various other Bending resistant systems include joist, beam, girder, as well as Vierendeel frame and
framing systems that resist gravity load as well as lateral wind- and seismic load.
1 9 9
girder, folded plate and cylindrical shell. They carry gravity load primarily in bending to a
support structure. Shear is typically concurrent with bending; yet bending usually
erle
In the interest of a structured presentation, both, horizontal and vertical/lateral systems controls the design. Though bending resistant elements and systems are very common,
i
are further classified by type of resistance controlling the design. This also helps to they tend to be less efficient than other systems, because bending varies from maximum
GS c h
structure the creative design process. Though many actual systems may include several
modes of resistance, the controlling resistance is assumed for the classification. For
example, cable stayed systems usually include bending elements like beams, in addition
compression to maximum tension on opposite faces, with zero stress at the neutral axis.
Hence only half the cross-section is actually used to full capacity. Yet, this disadvantage
G
is often compensated by the fact that most bending members are simple extrusions, but
t
to cables or other tension members. However, at least at the conceptual level, their trusses are assembled from many parts with costly connections. Like any structure
r g h
designed is controlled more by tension members than by bending. Therefore they are
i
classified as tensile structures. Horizontal systems are presented in four chapters for
y
system, bending elements are cost effective within a certain span range, usually up to a
maximum of 120ft (40m). For longer spans the extra cost of more complex systems is
Cop
structures controlled by bending, axial, form and tensile resistance. Vertical/lateral justified by greater efficiency.
systems are presented in three chapters for structures controlled by shear-, bending-,
and axial resistance.
PART IV HORIZONTAL SYSTEMS Bending Resistant

Bending Concepts
Some concepts are important for an intuitive understanding of bending members and
their efficient design. They include the effects of span and overhang, presented in this
section. Other concepts, such as optimization and the Gerber beam, are included in the
following section.
Effect of span
The effect of the span L for bending members may be demonstrated in the formulas for
deflection, bending moment and shear for the example of a simple beam under uniform
load.
Δ= (5/384) wL4/ (EI)
M= wL2/8
6
V= wL/2
where
Δ= Maximum deflection
0 - 2 0 0
9
E= Elastic modulus
1
I = Moment of Inertia
9
erle
L= Length of span
c h i M= maximum bending moment

V= maximum shear force
GS
w= Uniform load per unit length
i g h t G The formulas demonstrate deflection increases with the 4th power of span, the bending
moment increases with the second power, and shear increases linearly. Although this
r
example is for a simple beam, the same principle applies to other bending members as
Copy well. For a beam of constant cross-section, if the span is doubled deflection increases 16
times, the bending four times, but shear would only double. Thus, for long bending
members deflection usually governs; for medium span bending governs, yet for very
short ones, shear governs
1 Beam with deflection Δ = 1
2 Beam of double span with deflection Δ = 16
3 Short beam: shear governs
4 Medium-span beam: bending governs
5 Long-span beam: deflection governs
11-1 HORIZONTAL SYSTEMS Bending Resistant

Effect of overhang
Bending moments can be greatly reduced, using the effect of overhangs. This can be
describe on the example of a beam but applies also to other bending members of
horizontal, span subject to gravity load as well. For a beam subject to uniform load with
two overhangs, a ratio of overhangs to mid-span of 1:2.8 (or about 1/3) is optimal, with
equal positive and negative bending moments. This implies an efficient use of material
because if the beam has a constant size – which is most common – the beam is used to
full capacity on both, overhang and span. Compared to the same beam with supports at
both ends, the bending moment in a beam with two overhangs is about one sixth ! To a
lesser degree, a single overhang has a similar effect. Thus, taking advantage of
overhangs in a design may result in great savings and economy of resources.
1 Simple beam with end supports and uniform load
2 0 0 6
2 Cantilevers of about 1/3 the span equalize positive and negative bending
moments and reduces them to about one sixth, compared to a beam of equal length
-
0
and load with but with simple end support
1 9 9
c h i erle
h t G GS
y r i g
Cop

Bam optimization
Optimizing long-span girders can save scares resources. The following are a few
conceptual options to optimize girders. Optimization for a real project requires careful
evaluation of alternate options, considering interdisciplinary aspects along with purely
structural ones.
1 Moment diagram, stepped to reflect required resistance along girder
2 Steel girder with plates welded on top of flanges for increased resistance
3 Steel girder with plates welded below flanges for increased resistance
4 Reinforced concrete girder with reinforcing bars staggered as required
5 Girder of parabolic shape, following the bending moment distribution
1 Girder of tapered shape, approximating bending moment distribution
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Gerber beam
The Gerber beam is named after its inventor, Gerber, a German engineering professor at
Munich. The Gerber beam has hinges at inflection points to reduce bending moments,
takes advantage of continuity, and allows settlements without secondary stresses. The
Gerber beam was developed in response to failures, caused by unequal foundation
settlements in 19th century railroad bridges.
1 Simple beams over three spans
2 Reduced bending moment in continuous beam
3 Failure of continuous beam due to unequal foundation settlement, causing the span
to double and the moment to increase four times
4 Gerber beam with hinges at inflection points minimizes bending moments and avoids
failure due to unequal settlement
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Joist, Beam, Girder
Joists, beams, and girders can be arranged in three different configurations: joists
supported by columns or walls1; joists supported by beams that are supported by
columns2; and joists supported by beams that are supported by girders that are
supported by columns3. The relationship between joist, beam, and girder can be either
flush or layered framing. Flush framing, with top of joists, beams, and girders flush with
each other, requires less structural depth but may require additional depth for mechanical
systems. Layered framing allows the integration of mechanical systems; with main ducts
running between beams and secondary ducts between joists. Further, flush framing for
steel requires more complex joining, with joists welded or bolted into the side of beams to
support gravity load. Layered framing with joists on top of beams with simple connection
to prevent displacement only
6
2 Single layer framing: joists supported directly by walls
3
4
- 2 0 0
Double layer framing: joists supported by beams and beams by columns
Triple layer framing: joists supported by beams, beams by girders, and
0
9
girders by columns
5
1 9
Flush framing: top of joists and beams line up
erle
May require additional depth for mechanical ducts
6 Layered framing: joists rest on top of beams
c h i Simpler and less costly framing
GS
May have main ducts between beams, secondary ducts between joists
A Joists
i g h t G B Beam
C Girders
r
D Wall
Copy E Column
F Pilaster
G Concrete slab on corrugated steel deck

Crown Hall, IlT, Chicago (1956)
Architect: Mies van der Rohe
Crown Hall for architecture at the Illinois Institute of Technology, exemplifies Mies’
architecture of universal space and structural expression, exposing girders and columns
on the outside. His tectonic objectives of exposed girders above the roof reduces air
conditioned due to less interior volume but also implies penalties: the girder top is not
braced against buckling, and the roof is punctured at many suspension points for
potential leaks. A column-free space of 120x220ft (37x67m) is spanned by four moment
resistant steel portals of 14in (360mm) wide-flange columns and four ft (1.2m) deep plate
girders that span the entire width. The portals, spaced 60ft (18m), support steel joists
suspended from the girders on bracket hangers. The joists, spaced 10ft (3m) on center,
overhang 20ft (6m) at end portals. To resist buckling, the girders have stiffener plates
welded to the web at intervals of the joists. Besides stability, they give the girders a
6
tectonic articulation.
1 Wall cross section
2 Structural diagram
0 - 2 0 0
9
A Top flange of girder
B Column
1 9
erle
C Stiffener plate welded to girder web
c h i
D Cant-strip at roof edge
E Roofing membrane
GS
F Suspension brackets
G
G Roof joist
r i g h t H Ceiling
I Glass wall
y
J Concrete floor
Cop

National Gallery, Berlin (1968)
Architect: Mies van der Rohe
The National Gallery was initially commissioned in 1962 for Berlin’s twentieth century art
collection. In 1965 it was merged with the National-galerie and renamed accordingly. A
semi-subterranean podium structure of granite-paved concrete is the base for the main
structure; a steel space-frame of 64.8m (212ft) square has a clear interior height 8.4m
(28ft). At the roof edge eight cross-shaped steel columns with pin joint at the roof
cantilever from the podium for lateral resistance. Based on a planning module of 1.2m,
the unique space-frame consists of two-way steel shapes, 1.8m deep, spaced 3.6m on
center in both directions. The shallow span/depth ratio oh 33 required the roof to have a
camper to cancel deformation under gravity load. The entire roof was assembled on
ground from factory pre-welded units and hydraulically lifted in place on a single day.
6
1 Steel roof framing concept
2 Steel roof framing detail
A Steel edge beam
0 - 2 0 0
9
B Cross-shaped steel column
1 9
c h i erle
h t G GS
y r i g
Cop

School in Gurtweil, Germany (1972)
Architect: H. Schaudt
Engineer: Ingenieurbüro für Holzbau
A grid of equilateral triangles is the base of this honeycomb of ten hexagonal classrooms.
The side length of each regular hexagon is 5.4 m (18 ft). The composition of classrooms
defines a free-form hall with entry from a central court. The sloping site provides space
for a partial ground floor for auxiliary spaces below the classroom level.
Laminated beams spanning three ways presented a challenge to minimize the number of
beam intersections. The continuous beams need moment connections. To this end, the
main roof and floor structures have identical configurations but different support
conditions. Six columns support each hexagonal classroom at the vertices. The
6
classroom floors have an additional column at the center of each hexagon to support the
- 2 0 0
cross point of three girders that span the six hexagon vertices. Those columns do not
interfere with the auxiliary spaces below classrooms. Three beams span between the
0
9
girders to form four triangular panels. Floor joists rest on the beams and support a
1 9
particle board sub-floor with acoustical and thermal insulation. To provide uninterrupted
erle
classrooms, the roof structure has no column support within each hexagon. The column-
free spaces required beams with moment connections. The roof deck consists of planks
c h i
with tongue-and-groove. Diagonal steel rods, 24 mm (1 in) φ, with turnbuckles, brace
GS
some peripheral columns to resist lateral wind load.
1 Floor structure (roof is similar but without column at hexagon centers)
i g h t G 2
3
Column supporting center of floor hexagon
Moment resistant joint of roof beam at hexagon center without column
y r
Cop
A Laminated girder, 12x60 cm (5x24")
B Steel insert bar with dowels ties beams to column
C Hexagonal laminated column, φ 21 cm (8")
D High-strength concrete core resists compression at top of roof beam
E Steel strap, 10x80 mm (3/8x3"), resists tension at bottom of roof beam
E Tension straps, 10x80mm, at bottom of beams

Labor Palace Turin (1961)
Architect/Engineer: Pierre Luigi Nervi
This project, first price of a design competition, was build for the centenary of Italy’s
unification in 1981 to house an international labor exhibition. The classic order of this
structure is a departure from Nervi’s funicular oeuvre. Due to a short time from design to
completion, one of the design objectives was fast construction. The solution of 16
freestanding mushroom structures allowed for sequential manufacture and erection, a
critical factor for speedy completion. The facility measures 158x158m and has a height
of 23m. Each of the 16 units measures 38x38m. The mushrooms are separated by
gabled skylights of 2m width that help to accentuate individual units visually, provide
natural lighting and structural separation. Each mushroom consists of concrete pylons
6
that taper from 2.4m at the top to 5.5m at the base in response to the increasing
- 2 0 0
bending moment toward the base. The pylons are rounded at the top and cross-formed
at the base. Twenty tapered steel plate girders cantilever from the pylons in radial
0
9
patterns; with increasing depth toward the support in response to greater bending.
1 9
Triangular brackets strengthen the transition from girder to pylon. Stiffener plates welded
erle
to girder webs stabilizes them against buckling and provide a visual pattern in response
i
to the structural imperatives.
GS c h
i g h t G
y r
Cop

Vierendeel
Named after the Belgian builder Vierendeel who is credited with its invention, the
Vierendeel girder has rectangular panels, composed of struts, which are connected by
moment resistant joints. Since load is resisted in frame action rather than truss action,
the terms Vierendeel girder or frame (depending on configuration) are more suitable than
Vierendeel truss which implies triangular panels and axial stress rather than bending
stress. Compared to trusses with triangular panels, rectangular Vierendeel panels allow
ducts or pipes to pass through with greater ease. Rectangular panels also have a
different visual appearance. Vierendeel girders resist load in combined axial and
bending stress and, thus, tend to be less efficient than trusses of triangular panels, which
resist load in axial tension and compression. Bending stress in members varies from
zero at the neutral axis to maximum at the outer fibers, but axial stress is uniform and
thus more efficient. For convenience, we refer to horizontal and vertical struts as chord
2 0 0 6
and web, respectively. The load bearing of Vierendeel girders and frames can be
visualized by magnifying their deformation under load. A single-bay Vierendeel1 deforms
-
0
under gravity load similar to a moment resisting portal frame2: top and bottom chords
9 9
develop positive moments at mid span and negative moments at both ends, with two
1
inflection points at the transition. Chord rotation is transmitted to webs and deforms them
erle
into S-shapes. The resulting web moments are inverse on top and bottom, with inflection
i
points of zero moment at mid height. Under lateral load3 both chord and web struts are
GS c hdeformed with single inflection points in the middle. In multi-bay girders5, too, webs
deform under both gravity- and lateral loads similar to frames6-7, with inflection points that
G
may be hinged. However, the chords develop single inflection points for both lateral and
t
gravity loads, except the center bay which has two inflection points under gravity load. In
y r i g h girders with even number of bays and a center web strut, all chords have single inflection
points. Since all web struts, assume inflection points under both gravity and lateral load,
Cop
they could have hinges at those points, provided those hinges can resist out-of-plane
deformation to avoid instability.
1 Single-bay Vierendeel girder

2 Deformation under gravity load
3 Deformation under lateral load
4 Web struts with hinges at inflection points
5 3i-bay Vierendeel girder
6 Deformation under gravity load
7 Deformation under lateral load
8 3-bay web struts with hinges at inflection points

Vierendeel girders resist load in combed beam action and frame action as shown on the
left and right diagrams, respectively. Load causes global shear and bending which
elongates the bottom in tension and shortens the top in compression. The internal
reaction to global shear and bending is different in a Vierendeel compared to a beam. A
beam's bending stress varies gradually over the cross-section, but global bending in a
Vierendeel causes concentrated tension and compression forces in the chords. By visual
inspection we can derive simple formulas for approximate axial and shear forces and
bending moments. Respective stresses are found using formulas for axial, shear and
bending stress and superimposing them. Chord tension T and compression C are
computed, dividing the respective global moment M by frame depth D (distance between
centers of chords).
C=T=M/D
6
Bending of individual struts can be visualized too. In a structure where moment resistant
- 2 0 0
strut/chord connections are replaced with hinges, chords would deflect as independent
beams6. Assuming flexible chords and stiff webs, vertical shear would deform chords to
0
9
S-shapes with inflection point. Assuming flexible webs and stiff chords, horizontal shear,
1 9
caused by a compressive force pushing outward on top and a tensile force pulling inward
erle
on the bottom, would deform webs to S-shapes with inflection point. The combined effect
of these two idealized cases imparts S-shaped deformation and inflection points in both
c h i
chord and web struts. The deformation yields strut bending moments which vary from
GS
positive to negative along each strut. Top and bottom chords carry each about half the
total shear V. Assuming inflection points at midpoints of chords, the local chord moment
G
M is half the shear V multiplied by half the chord length.
rig h t M = (V / 2) (e / 2)
C o p y The moment M is maximum at supports where shear is greatest and equal to support
reactions. For equilibrium, webs have to balance chord moments at each joint. Their
moment equals the difference of adjacent chord moments.
1 Gravity load on a Vierendeel
2 Global shear (in overall system rather than individual members)
3 Global bending (in overall system rather than individual members)
4 Compression and tension in top and bottom chord, respectively
5 Free-body visualizes derivation of chord tension T and compression C
6 Global shear deformation
7 Chord bending, assuming flexible chords and stiff webs
8 Web bending, assuming flexible webs and stiff chords
9 Combined chord and web bending under actual condition
10 Free-bodies visualize derivation of chord bending moment M

Configurations
Vierendeels may have various configurations, including one-way and two-way spans.
One-way girders may be simply supported or continuous over more than two supports.
They may be planar or prismatic with triangular or square profile for improved lateral load
resistance. Some highway pedestrian bridges are of the latter type. A triangular cross-
section has added stability, inherent in triangular geometry. It could be integrated with
bands of skylights on top of girders.
When supports are provided on all sides, Vierendeel frames of two-way or three-way
spans are possible options. They require less depth, can carry more load, have less
deflection, and resist lateral load as well as gravity load. The two-way option is well
suited for orthogonal plans; the three-way option adapts better to plans based on
triangles, hexagons, or free-form variations thereof.
special case of the Vierendeel concept.
- 2 0 0 6
Moment resistant space frames for multi-story or high-rise buildings may be considered a
1
2
One-way planar Vierendeel girder
1 9 9 0
One-way prismatic Vierendeel girder of triangular cross-section
erle
3 One-way prismatic Vierendeel girder of square cross-section
4 Two-way Vierendeel space frame
c h
5
i Three-way Vierendeel space frame
GS
6 Multi-story Vierendeel space frame
i g h t G
y r
Cop

Beinecke library, Yale University, New Haven, Connecticut (1963)
Architect and Engineer: Skidmore, Owings and Merrill
The Beinecke library of Yale University for rare books has a 5-level central book tower,
freestanding within a single story donut-shaped hall that extends over the full height of
the tower. The tower holds 180,000 books and is climatically separated from the
surrounding hall by a glass curtain wall.
The library’s five-story open space is framed by a unique structural concept. Four
Vierendeel steel frames, 50 ft (15 m) high, support the roof and wall load and span 131
and 88 ft (40 and 27m) in length and width, respectively. The frames are supported by a
reinforced concrete plate that transfers the load via steel pin joints to four reinforced
concrete pylons. The Vierendeel frames consist of 8'-8" (2.6m) prefabricated steel
crosses, welded together during erection. The crosses express pin joints at mid-points of
6
chord and web struts, where inflection points of zero bending occur.
0 - 2 0 0
1 9 9
c h i erle
h t G GS
y r i g
Cop

Folded Plate
The effect of folding on folded plates can be visualized with a sheet of paper. A flat
paper deforms even under its own weight. Folding the paper adds strength and stiffness;
yet under heavy load the folds may buckle. To secure the folds at both ends increases
stability against buckling
1 Flat paper deforms under its own weight
2 Folding paper increases strength and stiffness
3 Paper buckling under heavy load
4 Secured ends help resist buckling
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Folded plate behavior Bending in folded plates causes top compression and bottom tension. Folded plates also
Folded plates combine slab action with beam action. In length direction they act like thin tend to flatten out under gravity load, which may be prevented by walls or frames at end
inclined beams of great depth, stabilized against buckling at top and bottom by adjoining supports. Tendency of end panel buckling can be resisted by edge beams.
plates. In width direction they are one-way slabs that span between adjacent plates.
1 Bending visualized as external compression and tension forces
1 Folded plate concept 2 Flattened folded plate under gravity load
2 Slab action in width direction 3 Folded plate with walls and frames to resist flattening
3 Slab-and-beam equivalent 4 Buckled end panels
4 Beam action in length direction 5 End panels stabilized by edge beams
A Bending deformation causes top compression and bottom tension
A Stabilizing wall at support and, for long systems, at mid-span
B Horizontal shear caused by compression and tension
B Stabilizing frame at support and, for long systems, at mid-span
C Vertical shear is maximum at supports and zero at mid span
C Edge beam to stabilize end panel against buckling
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Folded plate forms Folded plates may be two or three-way systems.
Folded plates may have many one-way, two-, or three-way spans. They may be
1 Three-way folded plate unit and assembly on triangular base plan
motivated by aesthetic or spatial objectives, or to add strength and stability to a system.
2 Two-way folded plate unit and assembly on square base plan
In areas with snow, flat folded plates are problematic since snow can accumulate in the
3 Three-way folded plate unit and assembly on hexagonal base plan
valleys. One-way systems are shown below; Two and three-way systems are right.
1 Folded plate with one straight and one gabled edge
2 Folded plate with offset gabled edges
3 Folded plate with gabled edges offset at mid-span
4 Folded plate with vertical support folding and gables offset at mid-span
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Railroad Station Savona, Italy (1958-61)
Architect: Antonio Nervi
Engineer: Pier Luigi Nervi
This first prize of a design competition consists of site-cast concrete folded plates,
supported by a folded plate concrete wall on the rear which also provides lateral stability.
Ten concrete pylons support the public entry front. The pylons are cantilevered from a
grade beam for lateral stability in length direction. The pylons transform from rectangular
cross-section in length direction at the ground to rectangular cross-section in width
direction at the roof. They support a u-shaped roof girder that is integral with and
supports the folded plate roof. One-third of the roof overhangs in front, beyond the
girder. The overhang is tapered, transforming from the folded plate profile to a flat roof
edge. The taper makes an elegant edge in logic response to the diminished negative
bending moment requiring less depth at the edge. Light-weight gabled roof elements
- 2 0 0 6
cover the folded plates over interior space for waterproofing. Over the central area
skylights, integrated in the roof, provide natural lighting.
1 Folded plate concrete roof layout
1 9
2 Typical folded plate concrete unit
9 0
erle
A Cross-section through roof overhang with tapered folded plates and u-shape girder
i
B Length-section through folded plates
GS c h
i g h t G
y r
Cop

Gunma Music Center, Takasaki, Japan (1961) Shopping Center, Würzburg, Germany
Architect: Antonin Raymond Architect: Schönewolff and Geisendörfer
Engineer: Tsuyashi Okamoto Engineer: Julius Natterer
This Gunma music center for the Gunma Philharmonic Orchestra, consist of a folded The folded plate wood roof modules are 7 m wide and span 16.25 m for the large space;
plate concrete roof of 60 m span and folded plate walls, that form frames to resist gravity 5 m wide and span 12.5 m for small spaces.
and lateral loads. The architect, a former student of Frank Lloyd Wright at Taliesin took 1 View of folded plate wood roof
the challenge to design the center for the following requirements: 2 Cross-section of typical folded plate module
• The center had to be fire and earthquake proof 3 Detail of valley joint
• Good acoustics for the music center
A Tie strut 135x520 mm
• Provide for Kabuki performances that required a revolving stage B Folded plate cross planking 4 cm
The folded plate roof is 3.3 m deep for a span/depth ratio of 1:18. Two wings flanking the C Transverse ribs, 8 x 16 cm, spaced 1.9 m
stage for meeting and green rooms, also have folded plate roofs.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Cylindrical Shell
The shape effect of cylindrical shells can be visualized with paper. A flat paper deforms
even under its own weight. To roll or bend paper into cylindrical shapes adds strength
and stiffness; yet heavy load may flatten and buckle the paper. Securing both ends
prevents buckling
1 Flat paper deforms under its own weight
2 Rolling paper increases strength and stiffness
3 Cylindrical form also increase strength and stiffness
4 Paper buckling under heavy load
5 Secured ends help resist buckling
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Cylindrical shell behavior
Considering their name, cylindrical shells could be part of shells; but they are included
here because they resist load primarily in bending, unlike shells which act primarily in
tension and compression. Most cylindrical shells have semi-cylindrical cross-sections
and act much like beams of such cross-section, spanning horizontally to transfer gravity
load to supports. Similar to beams under gravity load bending in cylindrical shells cause
compressive stress on top and tensile stress at the bottom; unlike vaults with primary
span in width direction. Differential bending stress, pushing and pulling on top and
bottom generates horizontal shear stress in cylindrical shells. To satisfy equilibrium,
horizontal shear causes also vertical shear which can be visualized as tendency of
individual parts to slide vertically with respect to one-another. Stress distribution over the
cross-section is also similar to beams. Bending stress varies from maximum
compression on top to maximum tension on the bottom, with zero stress at the neutral
6
axis. In contrast shear stress is maximum at the neutral axis and zero on top and bottom.
walls or ribs.
0 - 2 0 0
Compressive stress in cylindrical shells causes buckling which can be resisted by cross-
1
2
9 9
Compressive stress on top, tensile stress at bottom, with some arch action
1
Horizontal shear generated by differential compressive and tensile stress
erle
3 Vertical shear visualized
c h
4
5
i Bending stress distribution
Shear stress distribution
GS
6 Buckling under gravity load
G
7 Buckling under lateral load
r i g h t 8
9
Wall panels to resist buckling
Ribs to resist buckling
Copy

Configurations Skylights
Cylindrical shells can have various configurations: cross-sections of half or quarter Various skylight forms maybe integrated with cylindrical shells. This has been a popular
cylinders, or other curved forms; they may have closed ends or be open at one or both solution for natural lighting of industrial buildings. Combining the inherent strength,
ends; they may be simply supported, cantilevered, or span two supports with one or two stiffness, and stability of cylindrical shell forms with natural lighting is a logical design
overhangs. The end units may be open or closed. Butterfly cross-sections are also strategy. The skylights may be inclined in the shell form, or flat on top, or in the vertical
possible if designed to resist bending in width direction. The intersection between plane of a quarter-cylindrical shell. Skylights could be incorporated with a truss as part of
adjacent shells must incorporate a gutter to drain rainwater. In snow areas, horizontal the cylindrical shell. An important factor in integrated skylights is waterproofing to
cylindrical shells are problematic, since snow would accumulate in the valleys. prevent leaks, and to incorporate some form of gutter for drainage.
1 Semi-circular cylindrical shells, simply supported, with glass ends 1 Cylindrical shell with truss skylight
2 Shallow units cantilever from a beam, designed to resist rotation 2 Skylight on top of cylindrical shell
3 Butterfly units, cantilevered from pylons 3 Vertical skylight with cylindrical shell of quarter cross-section
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Kimbell Art Museum
Architect: Louis Kahn
Engineer: Kommendant
The Kimbell Art Museum is composed of three parts: the central main entrance, facing 1 Exploded isometric view
bookstore and library is flanked by two gallery wings, one on each side. The gallery 2 Cycloid, formed by a point on a cycle that moves horizontally
wings include atrium courtyards. The entire facility is composed of 16 modules of about 3 Cross-section of cycloid shells
30x100ft (9x31 m). The modules consist of cycloid shells, 24ft (7.3m) wide with a flat
A Point on the cycle that
part of 6ft (1.8m) between them (the cycloid cross-section is formed by a point on a
B Cycloid traced by the point on a cycle
moving wheel). A 30in (76 cm) wide skylight extends on top of each shell unit. A metal
C Linear skylight
deflector below each skylight reflects the daylight against the interior surface of the
D Reflectors of polished metal
cycloid shells for indirect natural lighting. The cycloid shells consist of post-tensioned
E Mechanical ducts
cast-in-place concrete. They where cast by using a movable form-work used repetitively.
F Duct cover
The flat roof between cycloid shells forms an inverted U to house mechanical ducts and
pipes as required.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

California Museum of Science and Industry
Architect: California State Architect Office
Engineer: T. Y. Lin and Associates
The roof of this rectangular museum consists of ten cylindrical shells and two half shells
as curved overhangs on the north and south sides. A group of eight inverted conical
shells provides a canopy for the main south side entry. The cylindrical shells provide
spatial relief and articulation for this stark rectangular plan. They continuo over two bays
and have span/depth ratios of 10. Post-tensioned tendons are draped to approximate a
parabola in space. Reflecting the bending deformation of the shells, the parabolic form
has an uplifting effect to counteract and minimize deflection. The tendons are
prestressed to produce a camber, designed to offset deflection due to dead load and
partial live load. The cylindrical shells where site-cast, using lightweight concrete (80% of
- 2 0 0 6
normal weight concrete) to minimize dead load. This is important in areas of seismic
activity, like Los Angeles, since seismic forces are proportional to mass, which
0
corresponds to deadweight. The shell thickness increases toward the base where they
9
form beams between adjacent units.
1 9
erle
1 Isometric roof plan
2 Length section in east-west direction
c h i
3 Typical shell cross-section
GS
A Post-tensioned prestress tendons, draped to offset deflection
i g h t G
y r
Cop

Kindergarten Yukari, Tokyo
Architect: Kenzo Tange
The fan-shaped plan of the Yukari Kindergarten for 280 children is designed in response
to a conic site of mild slope. The director of the facility, an artist, wanted an environment
of artistic inspiration for children of this kindergarten. The plan and space are strongly
defined by prefabricated cylindrical concrete shells, consisting of twin quarter-circular
elements with top stems for assembly and to hold the prestress tendons. Fan-shaped
shells accommodate the plan layout: each twin unit covers a modular space; large
spaces are covered by several units. Glass end walls emphasize the cylindrical shells
and extend them visually to the outside. Unit lengths vary with the spatial requirements.
The plan shows at left the roof and at right the floor plan with shells as dotted lines.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Tensile-resistant systems include stayed, suspended, cable truss, anticlastic, and
14 pneumatic structures. Although compression, bending and shear may be present in

some tensile structures, tensile stress is most prominent. For example, cable-stayed
systems may include bending resistant beams and joists, yet they are secondary to
primary stay cables or rods. Compared to bending and compression, tensile elements
are most efficient, using material to full capacity. Bending elements use only half the
material effectively, since bending stress varies from compression to tension, with zero
stress at the neutral axis. Compression elements are subject to buckling of reduced
capacity as slenderness increases. Furthermore, some tension elements, such as steel
H ORI Z ON T AL SY ST EM S cables, have much greater strength than columns or beams of mild steel, because they
are drawn (stretched) during manufacturing to increase strength. However, the overall
efficiency of tensile structures depends greatly on supports, such as ground anchors. If
T e nsile Re sist a nt
6
poorly integrated, they may require a large share of the budget. Therefore effective
- 2 0
anchorage is an important design factor. For example, the use of self-stabilizing
0
compression rings or infrastructures, such as grandstands, to resist tensile forces can be
0
9
an effective means of reducing support costs.
1 9
erle
Tension members
c h i1 Steel rod
GS
E = 30,000 ksi, Fa = 30 ksi, 100 % metallic
2 Strand consists of 7 or more wires (provides good stiffness, low flexibility)
ig h t G 3
E = 22,000 to 24,000 ksi; Fa = 70 ksi, 70% metallic
Wire rope consists of 7 strands (provides good flexibility, low stiffness)
r
E = 12,000 to 20,000 ksi, Fa = 70 ksi, 60% metallic
Co p y
1 2 3
14-1 HORIZONTAL SYSTEMS Tensile Resistant

Prestress
Tensile structures usually include flexible membranes and cables that effectively resist
tensile forces but get slack under compression. Yet, under some load conditions,
compressive forces may be induced in flexible tensile members. Prestress allows flexible
members to absorb compressive stress without getting slack which would cause
instability. Prestress also reduces deformation to half. These phenomena may be
observed on a simple string.
Consider a vertical string fastened on top and bottom. If a load is applied at mid-height,
the top link absorbs the entire load, and the lower link will get slack and unstable.
Now consider the same string prestressed (with turnbuckles for example). The same
6
load applied at mid-height will be carried half by the top link (through increase of
- 2 0 0
prestress) and half by the lower link (through decrease of prestress). Since both links are
active, each will absorb only half the load, reducing the deformation to half and avoiding
0
9
the lower link from getting slack and unstable. Since half the load is absorbed by each
1 9
link, when the applied load reaches twice the prestress or more, the lower link will get
erle
slack, just as the string with no prestress. Given similar conditions in a structure,
prestress should be at least half of the design load to prevent slack members and
c h i
instability. Also, loss of prestress due to creep and temperature variation should be
GS
considered.
G
The correlation between prestress, load, and deformation, described above, is visualized
r i g h t in the stress/strain diagram below.

1 String without prestress
Copy 2
3
A
B
String with prestress
Stress/strain diagram of both strings
Stress/strain line of un-prestressed string
Stress/strain line of prestressed string
C Point where prestress is reduced to zero under load
D Stress/strain line of string after loss of prestress
F Force
f Stress
P applied load
PS Prestress
Δ Deformation

Stayed Structures
Stayed structures consist of beams or trusses that are intermittently supported by strands
or rods (strands and rods have greater stiffness than wire ropes and hence reduce
deflection). Although stays usually support structures, pulling from above, they may also
push from below by means of compression struts. The latter is also referred to as cable-
propped or just propped. Given the slope of stays, they generate not only a vertical uplift
but also a horizontal reaction in the supported members and masts. In beams the
horizontal reactions yield compression; in masts they introduce bending and overturn
moments, unless stays on both sides balance the horizontal reactions.
The span/depth ratio of stayed and propped structures is an important design factor. A
shallow depth results in great tension and compression in stays and beam respectively, a
steep slope has the opposite effect. The relationship of cable slope and resulting forces
2 0 0 6
is illustrated in the diagrams, showing various slopes and resulting forces for an assumed
gravity load as vertical vector. Optimal span/depth ratios depend on both, architectural
-
0
and structural factors. Architectural factors include appearance and spatial
9 9
considerations. Structural factors include the impact on deflection, overall cost of stays,
1
beams, masts, and compression struts. As a rule of thumb, the optimal slope for stays is
erle
about 30 degrees. Optimum span/depth ratio for propped systems is about 10 to 15.
c h
1
i Steep stay slope causes small forces but high masts
GS
2 Stay slope of 25° to30° is usually optimal
3 Shallow stay slope causes high forces but low masts
i g h t G 4
5
Steep props cause small forces but great depth
Span/depth ratio of about 10 to 15 is optimal
r
6 Shallow props cause great forces but small depth
Copy

Configurations For light-weight roofs, with wind uplift greater than the roof dead weight, stays could by
Stayed structures may have radial or parallel strands, called radial and harp systems, added roof to resist wind uplift. Stays can also branch out like trees to reduce length.
respectively. Combinations of both systems are also possible. Harp systems have Single masts must be designed to resist overturning under unbalanced load. One-sided
constant stay forces; the force of radial systems varies with the stay slope. The tributary load causes unbalanced conditions that require guy cables. The dead weight of an
length between radial stays may be adjusted to keep forces constant, i.e., strands with inclined mast may help to balance loads.
shallow slope support small tributary lengths.
1 Stay cables below the roof resist wind uplift
1 Radial system (stay forces vary with slope) 2 Inverse tree stays reduce length but require more joints
2 Harp system (constant stay forces) 3 Single tower with tie-downs at both beam ends to resist overturning
3 Mixed system, combining radial and harp patterns 4 One-sided support with guy cable are unbalanced and less efficient;
4 System with variable distance between stay supports to equalize stay forces the inclined mast can help to balance the one-sided load
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

McCormick Place, Chicago (1987)
Architect: Skidmore, Owings and Merrill
Engineer: Knight and Associates
The expansion of McCormick Place exhibit hall, located over existing railroad tracks,
required a long-span roof to provide column-free exhibit space without interfering with the
tracks. Several structure systems had been investigated before selecting a stayed roof.
The roof is suspended from 12 concrete pylons, spaced 120x240ft, with 120ft overhangs
on both long sides. The pylons project 60 feet above the roof. The clear interior height is
40 feet. Stay cables consist of 3.75in galvanized steel strands, coated with corrosion
resistant PVC, arranged in parallel harp form at an angle of 25 degrees. The stays
support steel trusses which support secondary trusses, both 15ft deep and exposed at
the interior. The concrete pylons are shaped to incorporate mechanical ducts which
bring conditioned air from a mezzanine below the main floor and exhaust it over the roof,
2 0 0 6
without mechanical equipment exposed on the roof. The roof truss edges are tied to the
podium of the main hall to provide stability for unbalanced load. The podium is supported
-
0
by steel columns, spaced to accommodate the rail tracks. Combined with the deep
9 9
trusses, the stays have enough redundancy that they can be removed and replaced
1
without affecting the structure’s integrity. A glass band along the entire façade under the
erle
roof trusses and roof skylights, provide natural lighting.
c h1
i Isometric roof structure
GS
2 Cross-section of upper level with stayed roof
3 Mid-span stay support detail
h t G 4 Roof edge detail
rig
A Concrete pylons, shaped to accommodate mechanical ducts
C o p y B
C
D
E
Stays, 3.75in galvanized steel strands, PVC coated
Truss web bar
Stay connection bracket
Steel tie secures roof to [odium

Patscenter, Princeton, USA (1986)
Engineer: Ove Arup and Robert Silman
Patscenter is a research facility for PA Technology. The stayed roof structure was
chosen to provide column-free work space and to express technology as architectural
language desired by the client. One design criteria was to resist wind uplift load without
added dead weight on the roof. Based on a module of 4.5x9m, the single-story facility
measures 54x72m. The plan layout, as well as the structure and service technology are
all arranged along a central spine, 9m wide and flanked by two 22.5m wide work areas.
The entire roof is suspended from 9 triangular pylons, spaced 9m along the central spine
and supported by moment resistant portal frames. Steel stay rods on each side of the
pylons branch out to support the roof. Intersecting joints for the branches consist of
2 0 0 6
circular steel plates to which the rods are attached by means of standard fittings. Steel
rods where chosen over stay cables for greater stiffness and to facilitate painting. The
-
0
two inner stays are compression struts to resist wind uplift, with both outer stay secured
9 9
to columns that are tied to foundations. Using graphic vector analysis, the engineers
1
studied the geometry of the inverted tree branches to determine branch forces and
erle
overall stability. The roof rests on joists, spaced at 4.5m, spanning 9m between beams
c h i
which are suspended form the pylons by stay rods. The beams continue over the full
width of each wing. A platform for mechanical equipment is suspended by rods of
GS
triangular configuration to provide lateral stability for the pylons in length direction. In
G
width direction, lateral stability is provided by the triangular pylons and moment frames
r i g h t along the central spine.
Copy

Pan Am Terminal, J F K airport, New York (1959)
Architect/Engineer: Tippetts, Abbet, McCarthy, Stratton
This air terminal was designed with a large overhanging roof to protect boarding
passengers. Passenger circulation is straight forward. Departing passengers arrive at
the center and fan out to the peripheral departing gates. Arriving passengers proceed in
reversed direction. The structure, completed in 1959, was designed with an elliptical roof
of 422/528ft, with overhang that projects 114 ft beyond the building enclosure. Thirty-two
radial steel girders are supported by stays, each consisting of six 2.5in zinc-coated strand
bundles. The stays, attached near the edge of the steel girders, run over saddles of a
mid ring of concrete columns and are anchored to an inner ring of columns. The position
of columns is such that the girder load on both side is approximately balanced, a strategy
- 2 0 0 6
which made this giant overhang economically feasible. Steel joists, spaced 19ft span
between the radial girders to support the roof metal deck.
1
2
Roof plan
1
Section
9 9 0
erle
4 Mast top detail
S c h
A
B i Stays, 6 - 2.5 in φ strands at each
Stay saddle, rests on concrete columns
h t G G C 32 radial steel girders, 4.5ft to 7ft deep
p y ri g
Co

Propped Structures
Propped structures are supported from below rather than from above the horizontal span
members. They may consist of one or more struts which, propped by a strand or rod,
give elastic support to a girder. Struts require fixed connections to the girder to prevent
rotation; or they may form triangular configurations. The connection of tension members
may be concentric or eccentric. Concentric connections exert uniform compressive
stress on beams. Eccentric connections may be designed to cause negative support
moments that reduce the positive span moment and deflection.
A Concentric tie joint
B Eccentric joint (may reduce beam bending)
6
1 Twin struts with concentric tie connection
2
3
- 2 0 0
Single strut with eccentric tie connection to reduce beam bending
V-struts supporting two adjacent beams provide lateral bracing
0
9
4 Vertical and V-struts supporting three adjacent beams provide lateral bracing
5
1 9
Continuous propped beam
erle
6 Gable with propped rafters supported by buttress to resist lateral thrust
7 Gable with propped rafters and tie rod to resist lateral thrust
c h i
h t G GS
y r i g
Cop

St. Martin Church, Ingolstadt, Germany (1981)
Architect: Hempel and Brand
Engineer: Sailer + Stepan
A simple gable roof over a rectangular plan defines the space of this church. The
exposed wood structure adds natural warmth and a sense of balance. Six laminated
three-hinge twin-girder assemblies, span 20m across the full width of the nave. The twin
girders rest on concrete piers that cantilever from footings to resist gravity load, lateral
wind load and part of the outward roof thrust. Roof purlins that span between girders
support tongue-and-groove boards of diagonal patterns. The diagonal patterns stabilize
the roof for lateral wind load.
6
1 Structure system
2
A Roof purlins; 20x20cm, spaced 1m
0 - 2 0 0
Half section of three-hinge twin girders with prop cable and strut
B
C
9 9
Twin girders; 2-20x50cm, spaced 5m, span 20m
1
Steel rods support twin girders and resists outward thrust
c h i erle
h t G GS
y r i g
Cop

Concert Hall, Snape, UK (1967)
Architect/Engineer: Ove Arup
Remodeling this former malt house into a concert hall had to be done with care to
preserve the original character of the malt house, including the roof shape with four large
wood ventilation shafts. The concert hall with 840 seats measures 18.3x42m and has a
height of 15.5m to the flat part of the trapezoidal roof. The roof trusses are spaced 3.8m,
span 18.6 m and consist of:
• Twin rafters
• Compression struts
• Tension rods
The trusses are supported by peripheral walls. The twin rafters are propped by two
2 0 0 6
diagonal compression struts which are supported by tie-rods that are part of the lattice
truss. A compression strut links the rafters on top. Longitudinal joists support two layers
-
0
of planks that make up the roof diaphragm.
1 Roof plan
1 9 9
erle
2 Cross-section
3 Tension rod joint
c h i 4 Tension rod to rafter connection
GS
A Diagonal compression struts, 9.5x11cm lumber
G
B Steel plates
r i g h t C
D
Tie-rods, 19mm diameter
Top compression strut, 9.5x23cm lumber
y
E Rafters, 2 – 4.5x23cm lumber
Cop D

Suspended Structures
Suspended structures are used for long-span roofs. They are most effective if the
curvature is compatible with spatial design objectives, and the horizontal thrust is resisted
by a compression ring or by infrastructures, such as grandstands. Suspended cables
effectively resist gravity load in tension, but are unstable under wind uplift and uneven
loads. Under its own weight a cable assumes the funicular shape of a catenary (Latin for
chain line). Under load uniformly distributed horizontally, the funicular will be parabolic;
under point load the funicular is a polygon. Thus, without some means of stabilizing,
cables assume different shapes for each load. Furthermore, under wind uplift suspended
cables tend to flutter. Several means can be used to stabilize cables for variable loads
and wind uplift. Among them are stabilizing cables, anticlastic (saddle-shaped)
curvature, described later, and ballast weight. However, in seismic areas ballast weight
would increase the mass and thus lateral loads.
1
- 2 0 0 6
Suspended roof with compression ring to absorb lateral thrust
2
9 9 0
Suspended roof with grandstands to resist lateral thrust
1
erle
3 Catenary funicular under cable self weight
i
4 Parabolic funicular under horizontally distributed load
GS c h5
6
Polygon funicular under point load
Deformed roof under point load
ig h t G 7 Deformed roof under uneven load (snow at one side, for example)
p y r 8 Roof subject to wind uplift
Co
9 Roof with convex stabilizing cables to resist uplift and uneven loads
10 Dead load to resist uplift and reduce deformation under uneven load

Span/sag ratio
The span/sag ratio of suspended roofs is an important design factor (span is the
horizontal distance between supports and sag the vertical distance between supports
and cable low-point at mid-span). Considering constant gravity load, the effect of various
span/sag ratios can be seen by equilibrium vector polygons at the supports. Constant
gravity load causes approximately constant vertical reaction V for all sags, but horizontal
reaction H and cable tension T vary with the span/sag ratio. Consider the cable at left
under uniform load. The three equilibrium vector triangles below the cable clearly show:
• A small sag (shallow roof) causes a large cable force and horizontal thrust
• A big sag has the opposite effect but requires tal and more costly supports
The optimal span/sag ratio is usually about 10, depending on space requirements.
f
L
0 0 6
Sag: distance between supports and cable low point at mid-span
Length of span between supports
- 2
H
T
Horizontal support reaction
1 9
Maximum cable tension
9 0
erle
V Vertical support reaction
c h i
h t G GS
y r i g
Cop

Dulles airport terminal, Washington, DC (1958-62)
Architect: Ero Saarinen
Engineer: Ammann and Whitney
The Dulles international airport terminal near Washington, DC, has a cable roof
supported by concrete pylons. The outward leaning pylons partly resist the cable thrust.
Based on the dimensions of movable loading docks, designed by Saarinen, the pylons
are spaced at 40 ft (12m) for a column-free concourse space of 150x600ft (46x183m),
recently expanded, extruding the same structure. Given the slanted pylons, the
suspension cables actually span 161 ft (49 m). Concrete edge beams span the pylons at
heights that vary from 65ft (20m) along the entry to 40ft (12m) facing the runways.
Suspended from the edge beams are 128 bridge strands of ∅ 1in (25mm) which support
site-cast concrete roof panels. The concrete dead weight resists wind uplift and
minimizes roof deformations under unbalanced roof loads. In Saarinen’s own words the
2 0 0 6
Dulles roof is “a strong form between earth and sky that seems both to rise from the plain
and hover over it.” It presents functional integrity and synergy of form and structure.
-
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

1 Burgo factory, Mantua, Italy (1961-63)
Engineer: Pier Luigi Nervi
The large production machines of this paper factory required a column-free interior
space of 30x250m and a 140m opening between exterior supports. Nervi’s solution
was a roof structure like a suspension bridge. Two concrete frames support four
parabolic cables from which a flat concrete roof is suspended by hangers. The
frames are braced as inverted Y’s for lateral stability in length direction. The
suspension cables’ lateral thrust is resisted in the concrete roof slab. A glass wall
provides the non-structural enclosure.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS 2 Lufthansa aircraft hanger, Frankfurt (1968-72)
g
Architect: Beckert & Beckert
y r i Engineer: Helmut Bomhard, Dyckerhoff & Witmann
Cop
This aircraft maintenance hanger of 100x270m accommodates up to six 747 jets.
Large hanger doors required the roof to span the long way, with a concrete girder on
two columns supports at mid-span. Recessed columns create overhangs to reduce
the girder bending moment. The roof consists ten bands of pre-stressed, suspended
concrete slabs, separated by linear gable skylights. Given an overall height limit of
34m for air traffic safety, and an interior height of 24m, the roof structure was limited
to 10m depth for a span/depth ratio of 13.5 between supports. At both ends the
suspension roof rests on inclined supports with ballast weight to resist lateral trust.
Prismatic steel containers filled with concrete provide the ballast. Straight horizontal
tension strands resist outward support displacement under wind uplift, restrain the
ballast gravity load, and contribute to overall stability. Perpendicular struts tie the
suspended slabs together for rotational stability. The curvilinear roof, flooded with
natural light, creates a floating interior space, in contrast to the normally heavy
material of concrete.

Sports hall, Dijon, France (1976)
Architect: J. F. Devaliere
Engineer: R. Weisrock, SA
With floor plan dimensions of 47.25x70m, this sports hall has a suspension roof spanning
the length in response to an interior profile of spectator seating for 4,000. Glue laminated
tension girders; spaced 6.75m are suspended from concrete piers with pin joints. They
act primarily in tension, but have sufficient bending stiffness to resist deformation under
unbalanced gravity load and wind uplift. To facilitate transportation, they are spliced at
center. Glue laminated joists, spaced at 2.53m; support a metal roof with thermal
insulation. Wood struts brace the girders to the joists against rotation. Roof heating is
provided to remove snow. A grid of diagonal wood slats provides lateral wind bracing in
the roof plane.
1 Roof plan
- 2 0 0 6
2 Cross-section
9
3 Isometric roof framing detail
1 9 0
erle
A Glue laminated tension girders, 16x150cm
i
B Glue laminate joists, 11x33cm
GS c hC
D
Diagonal wood bracing slats
Girder bracing, 5x15cm
i g h t G
y r
Cop

Recycling hall, Vienna (1981)
Architect: L. M. Lang
Engineer: Natterer and Diettrich
This recycling center features a tent-like wood structure of 560 feet (170m) diameter that
soars to a height of 220 feet (67m) above ground, supported by a central concrete mast.
The suspended wood roof consists of 48 radial laminated ribs that rise from outer
concrete pylons with wood compression ring to the mast top. The ribs follow the funicular
tension line to carry uniform roof load in pure tension, but asymmetrical loads may cause
bending stress in the radial ribs that are designed as semi-rigid tension bands with some
bending resistance capacity. Diagonal boards form the roofing membrane and add shear
resistance to the assembly of ribs and ring beams. The cylindrical concrete support mast
cantilevers from a central foundation. It was designed to resist asymmetrical erection
- 2 0 0 6
loads and to contribute to lateral wind load resistance. The peripheral pylons are
triangular concrete walls with metal brackets on top to secure the radial ribs.
1
2
Cross section
9
Roof plan
1 9 0
erle
3 Top of central support mast
4 Typical roof assembly
c hA
i Radial laminated wood tension rib, 7.8x31-43 (20x80-110cm)
GS
B Laminated wood ring beams, 5x15in (12x39cm)
C Laminated wood compression ring
i g h t G D
E
Steel tension ring
Steel anchor bracket
y r
Cop

Cable Truss
Cable trusses evolved from needs to stabilize suspension structures against wind uplift
and unbalanced gravity loads, using a second set of cables with opposing curvature.
The Swedish engineer Jawerth developed a cable truss with diagonal brace cables
separating top support- and bottom stabilizing cables that resist wind uplift. This system
was widely used in the 1960’s. Lev Zetlin and other US engineers designed cable
trusses with various other configurations, including lintel shapes with compression struts
separating bottom support- and top stabilizing cables. In 1969 the author and his
students at UC Berkeley developed trusses with flat chord cables separated by
compression struts and diagonal truss cables. Model tests, a full scale prototype, and
extensive computer analysis demonstrated great stiffness of these trusses in one-way,
two-way, and three-way layouts.
6
1 Isometric of lintel trusses with bottom support- and top stabilizing cables separated
by vertical compression struts
- 2 0 0
2 Isometric of concave trusses with top supporting- and bottom stabilizing cables
0
9
separated by vertical tension struts
1 9
3 Lintel truss with diagonal compression braces
erle
4 Lintel truss with vertical compression struts
5 Concave truss with diagonal tension braces
c h i
6 Concave truss with vertical tension struts
GS
7 Concave/lintel truss with diagonal compression braces
8 Concave/lintel truss with vertical compression struts
i g h t G 9 Concave gable truss with fan support and stabilizing cables and central
compression strut
r
10 Concave gable truss with tension struts and central compression strut
Copy 11 Concave support cable and fan stabilizing cables

12 Parallel chord truss, vertical compression struts and diagonal tension braces

Parallel chord cable truss
The load bearing mode of parallel chord cable trusses is more complex than that of
concave or lintel type trusses since they have no funicular cable but it may be explained
as follows.
Consider a four-bay truss with loads P1 and P2. They are transferred to the supports be
a polygon formed by the center bay bottom chords and end-bay diagonal braces. A third
load applied at the center strut is transferred by a second polygon in conjunction with the
latter one. Thus, half the bars resist the load in active tension and the other (passive)
bars resist the load by reducing prestress. For uplift wind load the load bearing is
reversed with active bars becoming passive and vice versa. This load bearing mode
applies also to trusses with more bays, as long as they are prestressed in order for
passive bars to resist load by reducing prestress. In these trusses the prestress must be
externally stabilized. Trusses with compression chord bars may be internally stabilized
- 2 0 0 6
and simply supported. In order to avoid slack cables, prestress must be at least half of
the design load stress as described at the beginning of this chapter. When prestress
0
approaches zero under load, the bar forces are about equal to those in a conventional
1
static methods.
9 9
truss under equal load and proportions and can be found by graphic vectors or other
erle
1 Externally prestressed cable truss with four bays
c h
2
i Load bearing polygon formed to resist two loads
GS
3 Load bearing polygons to resist three loads
4 Externally stabilized truss with six bays
G
5 Internally stabilized truss with six bays
r i g h t
Copy

1 Open air theater Ötigheim, Germany (1961)
Architect: E. Heid
Engineer: Jawerth
The roof structure for this largest German outdoor theater resembles a hammock
with cable trusses that span between two girders that are supported by two columns
each. The cable trusses span 37 m (121 feet) between the girders and converge to
two ground anchors. The truss depth of 3.6 m (12 feet) results in a span/sag ratio of
about 20 for concave and convex cables that support gravity load and wind uplift,
respectively. Crescent-shaped seating layout, with two columns near the front edge,
provides unobstructed views for most seats. The prestressed cable trusses stabilize
the girders against buckling and rotation. The roof consists of metal deck and two
membranes over rigid insulation. The rigid insulation dampens rain water pounding,
rather than providing thermal insulation which is not needed for the outdoor theater.
2 Plan, open air theater Ötigheim
- 2 0 0 6
0
The plan shows the area between side girders represents as square of 37x37 m
1 9 9
(121x121 feet). The radial convergence of cable trusses toward ground anchors
induces compressive stress in addition to bending stress in the girders. The girder
erle
bending moments could have been greatly reduced by recessing the columns to
i
provide overhangs of about 1/3 the span between columns. The column recess
GS c h would have also improved unobstructed views from most seats.
i g h t G
y r
Cop 3 Factory at Lesjöers, Sweden
Architect: Lennart Bergström
Engineer: Jawerth
The factory features cable trusses, spaced 4 m (13 feet) with intermediate supports.
Five bays of 16 m (53 feet) and two end bays of 6 m (20 feet) provide a total length of
92 m (302 feet). Continued arrangement balances lateral trust of adjacent trusses.
Linear skylights over the supports and at truss mid-spans provide natural lighting.
The inclined end supports equalize forces in guy cables and truss cables. The angle
of inclination can be determined by graphic vectors: equal angles between mast and
cables causes imply equal cable forces.

Auditorium Utica, USA (1958)
Architect: Gehron and Seltzer
Engineer: Lev Zetlin
For a seating capacity of 6,500, the Utica auditorium has a circular plan of 240 ft (73m)
diameter. Radial cable trusses of concave lintel profile provide the roof structure.
Compression struts separate bottom and top strands. The cable trusses are supported
by a circular exterior concrete compression ring and connected to two steel tension rings
at the center. The circular compression ring is a highly efficient method to support cable
roofs, eliminating the need for costly lateral supports. The cable trusses are prestressed
by jacking the central tension rings apart. Different lengths of top and bottom chords
induced different prestress and natural frequencies to reduce vibration due to wind gusts.
The bottom support cables are 2in (50mm) zinc coated strands with 175kip prestress.
- 2 0 0 6
The stabilizing top cables are 15/8in (41mm) strands with 135kip prestress. The roof was
erected in three weeks, with temporary support of the central tension ring only.
A
B
Circular concrete compression ring
1 9 9 0
Top stabilizing cable, 15/8” (42mm) strands
erle
C Steel compression struts
i
D Bottom cable, 2” (50mm) strands
GS c hE Steel tension ring
i g h t G
y r Convex alternate
Cop
It is of interest to consider the implications of an alternate convex roof. Cable trusses of
convex profile require two outer compression rings but only one central tension ring. The
vertical compression struts can be replaced by tension rods. Two compression rings
would likely cost more than a single compression ring since compression rings have to
be designed to resist buckling under unbalanced load... The inward sloping roof would
require rain water to be removed by pumps, but the concave Utica roof is self-draining.

Anticlastic Structures
Anticlastic tensile structures are flexible membranes or cable nets of saddle-shaped
curvature. The term membrane is used here to imply membranes and cable nets. Given
the nature of flexible membranes, double curvature and prestress are essential for
stability. This can be observed with simple string models. Two strings pulled in opposite
directions stabilize a point at their intersection. If the strings are in non-parallel planes
the stability will be three-dimensional. Similarly, if a series of strings cross in opposite
directions they stabilize a series of points at their intersection. The cross points form a
surface, stabilized by anticlastic curvature. The surface may be a membrane of fabric or
other material or a cable net. Although anticlastic curvature provides stability, some
elastic deformation is possible due to material elasticity. Thus, steel cable nets with high
elastic modulus deform less than fabric membranes with lower stiffness.
6
In addition to curvature, prestress is also required to stabilize anticlastic membranes.
- 2 0 0
This too can be observed on a string model. Applied load elongates one string in tension
and shortens the other in compression. Without prestress, the compressed string will
0
9
get slack and unstable; but prestressed strings absorb compressive stress by reduction
1 9
of prestress. Since prestress renders both strings active to resist load, the resulting
erle
deflection is reduced to half compared to non-prestressed condition where only one
string is active. This observation is also described under Prestress at the beginning of
c h i
this chapter.
GS
Flat membranes are unstable. This, too, can be observed on a string model. Two
strings in a flat surface must deform into a polygon to resist load (a straight string would
i g h t G assume infinite forces). Therefore, flat membranes are unstable under load. Similarly,
synclastic (dome-shape) membranes would deform excessively under gravity load and
y r flutter in wind.
Cop
1 Two strings crossing in non-parallel planes stabilize a point in space
2 A series if strings (or a membrane) form a stable surface
3 Without prestress, one string (or series of strings) would get slack under
load, causing instability
4 Strings in a flat surface deform excessively under load, causing instability

Minimal surface Minimal surface equations
As the name implies, a minimal surface covers any boundary with a minimum of surface The minimal surface models also revealed equations that define the principle curvature of
area. The minimal surface is defined by three criteria: equilateral minimal surfaces (Schierle, 1977):
• Minimum surface are between any boundary Y= f1(X/S1) (f1+f2)/f1 + X tan φ
• Equal and opposite curvature at any point Y= f2 (Z/S2) (f1+f2)/f2
• Uniform stress throughout the surface The equations are based on empirical studies of minimal surface models of plastic film,
A minimal surface may be anticlastic or flat. A surface of flat or triangular boundaries is measured by means of a projected light grid with an accuracy of only 1.26% standard
always flat. Flat membranes are unstable structures. Increased curvature increases deviation. The findings were first published in the Journal for Optimization Theory and
stability. The minimal surface can be studied on soap film models; but they disappear Application (Schierle, 1977)
quickly. The author studied quadrilateral plastic models that keep a minimal surface after Although the equations are for minimal surface of quadrilateral plans, they provide
drying. The models revealed: reasonable accuracy for other boundaries as well. This should be further studied.
f1/f2 = A/B 1 Plan view of quadrilateral minimal surface
This is contrary to Hyperbolic Paraboloid shells. The surface of HP shells passes at mid- 2 Length section of quadrilateral minimal surface
6
height between low and high points regardless of boundary conditions. 3 Length section with Y-axis vertical
1 Minimal surface of square plan
2 Minimal surface of rhomboid plan
4 Cross section
0 - 2 0 0
9
3 Hyperbolic Paraboloid of square plan
4 Hyperbolic Paraboloid of rhomboid plan
1 9
c h i erle
h t G GS
y r i g
Cop

Cable nets
Anticlastic cable nets may approximate a minimal surface. However, even if mesh
cables have equal prestress, minor mesh distance variations cause uneven density and
stress distribution. Various cable net configurations are possible. The method of
manufacture is a primary factor in defining a cable net. If assembled on-site, support
cables may be strung from edges to hang in vertical planes. Stabilizing cables may also
be placed in a vertical plane position. Such a net could have perfectly square meshes in
plan projection. Cables could also be arranged following the principle curvatures with the
least distance across the surface. Such a net is more complex to manufacture. A cable
net can also be prefabricated as an orthogonal grid of equal meshes flat on the ground.
The meshes deform from squares into rhomboids to assume the anticlastic curvature.
This method was used for the German pavilion of the 1968 Montreal Worlds fair and for
sports facilities of the 1972 Olympic Games in Munich. Cables parallel to the edges are
6
nearly straight, like generating lines of a Hyperbolic Paraboloid shell. Such flat cables
- 2 0 0
will deform much more under applied load than curved cables. This was first
demonstrated by tests the author conducted with students at the University of California,
0
9
Berkeley in 1967. One cable nets was tested with curved cables and the other with flat
1 9
cables. Under equal load the net with flat cables deformed about six times more than the
erle
one with curved cables. The test results, drawn as dots on both nets, show the great
difference in stiffness. It is clear from this test, later confirmed by computer, that nets
c h i
with flat cables are not a viable solution as structures. The test results have been widely
GS
published.
G
1 Net with cables arranged in vertical planes, with square grid projection
rig h t 2 Net with cables in direction of principle curvature
y
3 Net of square meshes, prefabricated flat, deform into rhomboids in space
Co p 4
5
6
Net of cables running nearly straight in direction of generating lines
Net 3 under gravity load (dots show small deformations)
Net 4 under gravity load (dots show 6 times greater deflection than net 3)

Edge conditions Surface conditions
Anticlastic structures may have three edge types: cable, arch, beam/truss. Cables Anticlastic membranes may have four surface conditions with many variations: saddle
resist membrane stress in tension, forming a curve of equal radius in space under shape, wave shape, arch shape, point shape. The most appropriate type for a given
uniform prestress. Arches resist primarily in compression, with bending under some situation is determined by architectural and structural considerations.
loads. Edge beams resist in bending and trusses axial mode. Regarding architectural
objectives, straight beams or trusses are easy to connect walls, but edge cables require
more complex enclosures.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Pioneering structures (1953)
State Fair Building, Raleigh, North Carolina

Architect: Novicki and Deitrick
Engineer: Severud, Elstad, Krueger
Built in 1953, the Raleigh arena was the first saddle shape roof of curved cables
(Schuchow’s 1889 Nischni Nowgorod exhibit anticlastic cable roof had straight cables.
The support cables span 300 ft with 31 ft sag, for a span/sag ratio of 9.7. The concave
support cables, spaced about 6 ft vary from ¾ to 15/16 inch, the convex cables vary from
½ to ¾ inch. The concrete compression arches are inclined about 22 degree to the
horizontal and supported by concrete covered steel columns. The roof consists of
corrugated metal with thermal insulation.
- 2 0 0 6
1 9 9 0
erle
Hockey rink, Yale University (1958)
Architect: Ero Saarinen
c h i
GS
The 200/85 feet rink, built 1958, has an arch supported anticlastic cable net. The central
concrete arch is designed to resists unbalanced load in bending, rather than using the
i g h t G cable net for stability. Both support and stabilizing cables are 1¾ inch diameter, spaced
6 feet. The roof consists of 2x8 inch wood planking, nailed to 2x6 inch wood strips. The
y r oval plan provides most spectators seating at the preferred location near the rink center.
Cop
Entry arch Kőln (1957)
Architect: Frei Otto
Engineer: Fritz Leonhard
This arch marked the entry of the 1957 Kőln garden show. The steel arch of 112 feet
span and only 7.5 inch diameter is stabilized by the membrane to resist unbalanced
loads and wind uplift. The membrane, projecting 39 feet on both sides of the arch is
prestressed by edge cables which are supported by four steel masts.

Wave shapes
Recycling center project, Mil Valley, California (1971)
Architect: Neil Smith and G G Schierle
Engineer: G G Schierle
This structure was designed to expand easily in response to needed growth, using a
modular system. A base module of 25x80ft (7.6x24m) is supported by 22ft (6.7m) high
masts. Two half end modules provide enclosure at both ends. Only one base module
with both end enclosures is shown here. For sustainable energy efficiency the
membrane was designed of translucent natural canvass allowing natural daylight. Edge,
ridge, and valley cables where designed as bridge rope for flexibility in adjusting to the
curvatures. Membrane prestress was introduced by turnbuckle adjustment at cable ends
anchored to helix ground anchors. Variable prestress was required in order for the ridge
2 0 0 6
cables to remain in vertical plains as required for repeatability of the modules. The
prestress levels were determined by computer analysis. Mats were designed as
-
0
standard steel pipes with pin joint attachment to the foundation. The pin joints avoided
9 9
bending stress for optimal efficiency; moment resistant joints would introduce bending
1
stress in the masts under any movement.
c h i erle
h t G GS
y r i g
Cop

San Diego Convention Center (1987-89)
Architect: Arthur Erickson and a joint venture team
Engineer/membrane designer: Horst Berger
The San Diego Convention Center features a linear plan of 1.7 million square feet
(157,935 m2). Part of the top level is designed as semi-outdoor exhibit space, covered
with a wave-shape membrane roof covering an area of 300 x 300 feet. The membrane
undulates between ridge and valley cables that are suspended from triangular concrete
pylons spaced 60 x 300 feet. Openings at membrane ridges provide natural ventilation.
The openings are protected by secondary membranes hovering above. Flying buttress
masts, supported by guy cables, hold up the ridge cables to provide a column-free space.
The guy cables are also suspended from the triangular concrete pylons. This creative
support system makes the translucent roof appear hovering seemingly weightless above
the space which is flooded in filtered natural light. The lightweight roof provides a stark
- 2 0 0 6
contrast to the conventional concrete infrastructure. The concrete pylons reinforce this
contrast with compelling elegance. The Teflon-coated glass fiber membrane provides a
0
fireproof enclosure as required for permanent structures.
1 9 9
c h i erle
h t G GS
y r i g
Cop

Schlumberger Research Center, Cambridge, UK (1984)
Architect: Michael Hopkins
Engineer: Ove Arup
The Schlumberger Center conducts basic research for oil exploration which includes
drilling, fluid mechanics, etc. that work in close cooperation. Thus a major objective was
to facilitate contacts among theoretical and experimental researchers as well as
administrators. The client also required an option for future expansion. The design
features a 24m wide central space for drilling equipment and a recreation area between
two single story research wings with private offices, discussion rooms and laboratories,
all separated from the central space by 21mm thick sound insulation glass. The central
space is covered by a removable fabric, suspended from a network of cables that are
supported by braced steel frames. Three wave-shape membranes have two ridge cables
each. The ridge cables are suspended from overhead guy cables. The translucent fabric
equipment.
- 2 0 0 6
provides natural lighting and is removable to provide access for replacing the drilling
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Skating rink, Munich (1983)
Architect: Kurt Ackermann
Engineer: Sclaich Bergermann
This facility, initially designed as ice skating rink was recently converted into an inline-
skating facility due to the increasing popularity of this new sport. The elliptical rink of
88x67m is covered by a cable net roof, suspended from a central arch and supported
along the edges by a series if steel masts with guy cables. A prismatic trussed steel arch
spans 104m between concrete abutments. The arch supports the cable net and is itself
stabilized by it. The cable net is suspended to the arch by means of looping edge cables
along the central spine. The space between the edge cables is designed as a skylight
that exposes the arch from the inside and provides natural lighting in addition to a
translucent roofing membrane. The cable net of double strands has 75x75cm meshes to
which a lattice grid of wood slats is attached at the joints. A translucent PVC membrane
- 2 0 0 6
is nailed to the lattice grid. This unusual combination of materials creates a unique
interior spatial quality of quite elegance, contrasting the lightness of the translucent fabric
0
membrane with the warms of the wood lattice grid.
1 9 9
c h i erle
h t G GS
y r i g
Cop

Garden show pavilion Hamburg (1963) Ice skating rink Villars, Switzerland (1959)
Architect: Frei Otto Architect: Frei Otto
This pavilion for the international Garden Show 1961 covers an area of 29/64m and has The sunken skating area is surrounded by spectator seating and covered by a point
5.5m high masts. The point shape roof was fabricated as flat fabric without patterns. shape canvass membrane roof of 32x64m. The roof membrane hangs from three
The canvass stretched enough to assume the curvature between high and low points. suspension cables that span the length of the rink with steel masts at both ends. Metal
The high points are supported by steel masts with laminated wood springs over dishes distribute the membrane stress at support points. Light fixtures are suspended
octagonal steel ring to avoid stress peaks. Low points are anchored to the ground to from the same support points. Guy cables anchor membrane edge cables to the ground.
resist wind uplift and act as drainage points with rain water collector basins. Membrane
edge cables are anchored to the ground by guy cables.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Pneumatic Structures
Pneumatic structures are flexible membranes that derive their stability from air pressure.
They usually have synclastic curvature like domes, but anticlastic curvatures are possible
as well. Two generic types of pneumatic structure are air supported (low pressure) and
air inflated (high pressure) systems. The air pressure in inflated high pressure structures
is 100 to 1000 times greater than in air supported low pressure structures.
Air supported structures typically have a single fabric layer enclosing a space in form of
domes or similar shapes. The fabric is supported by inside air pressure. However,
considering human comfort, air pressure can be only slightly higher than outside
atmospheric pressure. The low air pressure makes air supported structures more
vulnerable to flutter under wind load. Since the usable space is under air pressure, door
openings must have air locks, usually in form of revolving doors to minimize loss of air
6
pressure. Air supported structures require continuous air supply, usually with standby
0 - 2 0 0
electric power generator to retain air pressure in case of power outage.
Air inflated structures are hermetically enclosed volumes that are inflated under high
9 9
pressure much like a football to provide stability. They can have various tubular or
1
cushion forms with high air pressure between two layers of fabric that provide usable
erle
space under normal air pressure. The air pressure ranges from 2 to 70 meters of water,
i
yielding 2.8 to 100 pounds per square inch pressure, enough to resist gravity and lateral
GS c hload. Without air pressure they would have no stability. Air inflated structures also
require some continuous air supply to make up for pressure loss due to membrane leaks.
i g h t G 1
2
Air supported dome or vault
Air supported vault
r
3 Air supported vault with support cables
Copy 4
5
6
7
Air supported dome repetitions
Air inflated cushion
Air inflated tubular vault
Air inflated tubular dome
8 Air inflated cushion repetitions

US Pavilion, Expo 70, Osaka (1970)
Architect: Davis, Brody, Chermayeff, Geismar, De Harak
Engineer: David Geiger
The US Pavilion was the first large-scale pneumatic structure in 1970 with an elliptical
plan of 466x272 feet (142x83 meters); yet rising only 20 feet (6 meters) from a peripheral
earth berm, the structure had a very low profile. This shallow curvature was possible
because the translucent roof membrane was laced to a grid of diagonal cables, spaced
20x20 feet (6x6 meter) that provided the primary support. The tension cables were
supported by a concrete compression ring on top of the earth berm by means of
adjustable anchor bolts. The compression ring formed a gutter to collect rain water along
the periphery. Bending moments that could have been generated in the compression
ring resulting of asymmetrical loads, were transferred to and resisted by the earth berm.
The pavilion impressed not only by its great size but by its combination of
1 Elliptical roof plan
- 2 0 6
understatement and technical innovation and refined sophistication.
0
2
3
Length section
9
Laced membrane to cable attachment
1 9 0
erle
4 Concrete compression wing with gutter and adjustable cable anchors
c h i
h t G GS
y r i g
Cop

Fuji Pavilion, Expo 70, Osaka (1970)
Architect: Yutaka Murata
Engineer: Mamoru Kawaguchi
The Fuji Pavilion housed an exhibit and light show of the Fuji Corporation in a unique,
organic form. Over a circular floor plan of 164 feet (50 meter) diameter, the pavilion
featured a vaulted fabric structure composed of 16 pneumatic arched tubes. The tubes
of 13 feet (4 meter) diameter were tied together by horizontal belts at 13 feet (4 meter)
intervals. The tubes consisted of two vinyl fabric layers that were glued together for
improved tear resistance. Given the circular floor plane, the arching tubes of equal
length form cross sections that vary from semi-circular at the center to semi-elliptical at
the entries on both opposite ends. To adjust the structure’s stiffness in response to
various wind pressures, the tubes were connected by pipes to a multi-stage turbo blower
that provided 1,000 to 2,500 mm water pressure.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Atoms for Peace pavilion (1960)
Architect: Victor Lundy
This pavilion housed a traveling exhibit of the United States Atomic Energy Commission
that was sent throughout Central and South America in 1960 when the adverse effect of
atomic energy were not yet fully understood. The pavilion included a cinema with seating
for 300 and a demonstration reactor under a double skin air supported structure. The
structure had air pressure between the two fabric layers as well as the inside usable
space; seemingly combining air supported and air inflated technologies; but it is indeed
air supported. The double skin fabric of vinyl coated polyamide improved the thermal
performance and added structural rigidity. The space between the double membrane
- 2 0 0 6
was divided into air chambers, separated by fabric ribs that provided additional strength.
The stout outside form reflected the interior space, given the constant spacing of 4 feet
0
(1.2m) between the two membranes. The portable pavilion of 131/328 feet (40x100
1 9 9
meters) was erected at each new exhibit site in 12 days by a crew of 12 workers.
c h i erle
h t G GS
y r i g
Cop

PART V 15
V ERT I CAL SY ST EM S V ERT I CAL SY ST EM S
Ge ne ra l Ba c k ground
- 2 0 0 6
Vertical structures are presented in four categories, considering primary resistance to
load: shear, bending, axial, and suspended (tensile). Although most structures combine
1 9 9 0
Vertical structures have been a challenge since the famed tower of Babylon. Motivations
to build tall structures include: a desire to reach toward heaven; to see the world from
erle
several categories, one usually dominates. For example, axially stressed braced frames above; the prestige of being tallest, and high land costs. The tallest church tower in Ulm,
i
Germany exemplifies the spiritual motivation. The Eiffel tower allows seeing Paris from
h
may also have moment resistant joints, yet the bracing provided most strength and
c
stiffness. above. The towers of the Italian hill-town San Gimignano, and contemporary corporate
GS
office buildings express power and wealth; the latter are also motivated by high land cost.
G
Traditional building materials like wood and masonry imposed height limitations
r i g h t overcome by new materials like steel and prestressed concrete. The Eiffel tower in Paris
marks the beginning of tall steel towers. Prestressed concrete towers were pioneered
y
1955 by Fritz Leonhard with a television tower in Stuttgart.
Cop
15-1 VERTICAL STRUCTURES General Background

1 SDR television tower Stuttgart 1955 217 m 712 feet
Tall Structures
2 Eiffel tower Paris 1889 300 m 984 feet
3 Chrysler building New York, 1930 319 m 1047 feet
4 John Hancock tower Chicago 1968 344 m 1127 feet
5 Bank of China tower Hong Kong 1988 369 m 1211 feet
6 Empire State building New York 1933 381 m 1250 feet
7 Sears tower Chicago 1974 443 m 1453 feet
8 CN tower Toronto 1976 553 m 1814 feet
9 Transmission tower Warsaw 1974
- 2 0 0 6 643 m 2110 feet
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Gravity load
Gravity load is the combined live and dead load, acting vertically to generate
compressive stress in supporting columns or walls. At every level they carry the
combined loads from above. Since load accumulates from top down, members at the top
carry the least; those at the bottom carry the most. Steel structures usually have the
same nominal column size but of increasing unit weight, resulting in thin columns at top
6
and bulky ones at ground level. For example W14 wide flange columns come in many
- 2 0 0
weights from 43 to 730 plf (64 to 1,086 kg/m) with capacities of 272 to 4,644k (1,210 to
20,656 kN). It is also possible to use higher strength steel at lower floors. However,
0
9
increase in steel strength does not yield higher stiffness since the modulus of elasticity of
1 9
steel is constant regardless of strength. For concrete structures it is possible to increase
erle
concrete strength and stiffness, or to increase the cross sectional area. If a mechanical
room is on the top floor it is possible to balance the decreasing need for duct sizes from
c h i
top down, with need for increasing column sizes from top to bottom. Eero Saarinen
GS
designed the CBS tower New York with such a strategy but was only partly consistent
since the lower floors are served from a mechanical room on the second floor.
i g h t G 1
2
structure with increasing column size as load increases from top down
Light-weight wide-flange column
y r 3 Heavy weight wide-flange column of equal nominal size as in 2 above
Cop
4 Increasing column size dovetails with reducing duct size from top down
5 Small column cross section and large duct size on top column
6 Large column cross section and small duct size on lower floor column
7 Large column on ground level where no duct space is needed

Lateral load
The effect of lateral load on tall structures is similar to gravity load on cantilevers, such as
balconies. Tall structures act like cantilevers projecting from the ground. Lateral load
generates shear and bending that may be presented in respective shear and bending
diagrams as in a cantilever beam. Yet there are important differences. The shear and
bending diagrams for buildings are usually global, for the entire system rather than for
individual elements like beams. For example, global bending (overturn moment) causes
axial tension and compression in columns, and local shear and bending in beams.
Further, lateral wind and seismic loads are non-uniform. Wind force increases with
height due to higher wind speed and reduced friction. Seismic forces increase with
height in proportion to increasing acceleration (acceleration increases with height due to
increased drift). However, shear increases from top to bottom since the structure at each
floor must resist not only the force at that floor but the forces from all floors above as well.
- 2 0
Cantilever beam with shear and bending deformation6
The non-uniform wind and seismic loads cause nonlinear shear distribution.
0
2
3
1 9 9 0
Tall building with shear and bending deformation
Shear and bending diagrams for uniform load on a cantilever beam
erle
4 Shear and bending diagrams for idealized uniform load on a building
5 Vertical distribution of wind force, shear, and bending diagrams
c h
6
i Vertical distribution of seismic force, shear, and bending diagrams
GS
A Load/force diagram
G
B Shear deformation
t
C Bending deformation
y r i g h D
E
Shear diagram
Moment diagram
Cop

Lateral resistance
Lateral loads may be resisted by shear walls, cantilevers, moment frames, braced
frames, or combinations thereof. The choice of a suitable system depends on structural
and architectural objectives. Shear walls and braced frames are strong and stiff,
cantilevers and moment frames are more ductile, to dissipate seismic energy. Shear
walls are good for apartments or hotels that require party-walls between units. Moment
frames offer better planning flexibility required for office buildings with changing tenant
needs.
Shear walls resist lateral load primarily through in-plane shear. They may be of
reinforced concrete or masonry, or, for low-rise, of wood studs with plywood or particle
board sheathing. Short shear walls tend to overturn and must be stabilized by dead load
or tie-downs.
6
Cantilevers are slender elements that resist load primarily in bending. Pole houses are
- 2 0 0
cantilevers; but more commonly, cantilevers are of reinforced concrete or masonry,
anchored to foundations, wide enough to resist overturn. Overturn cause compression
0
9
on one side and tension on the other. Compression acts in addition to gravity, tension
1 9
may be partly offset by gravity compression. In tall cantilevers, tension due to lateral load
erle
may be greater than gravity compression, resulting in net tension.
c h i
Moment frames consistent of posts and beams connected by moment resistant joints.
They may be of steel or reinforced concrete. To resist seismic load, concrete should
GS
have ductile reinforcement that yields before brittle concrete failure. Ductile design
G
results in greater concrete members with less reinforcing steel.
rig h t Braced frames may have diagonal-, A-, X-, or V-braces. The best bracing scheme
depends on structural and architectural considerations. K-bracing tends to buckle
Co p y columns and must not be used. X-bracing allows no doors and requires more joints for
greater cost; but X-bracing can be of tension rods to eliminate buckling. A- and V-braces
are shorter than single diagonals and result in reduced buckling. (However, beams must
be designed for the full span since bracing may adversely affect the beam load). Braced
frames are usually of steel but may be of reinforced concrete or wood (for low-rise).
1 Long shear wall resists in-plane load in shear primarily
2 Shear wall supports adjacent bays (slender walls tend to overturn)
3 Cantilever resists lateral load primarily in bending
4 Cantilever supporting adjacent bays
5 Moment frame requires moment resisting beam-column joints to resist lateral load
by beam-column interaction
6 Moment frames at both ends supports intermittent bays
7 Braced frame with diagonal bracing
8 braced center core supports adjacent bays

Braced /moment frame
Combined braced/moment frames are used to reduce drift under lateral load. Moment
frames have the greatest drift at the building base, but braced frames have the greatest
drift on top. Combining the two systems reduces drift at both base and top. The
objectives to reduce drift are:
• To prevent occupant discomfort

• To reduce the risk of failure of cladding and curtain walls
• To reduce secondary stress caused by P-Δ effects
(the P-Δ effect generates bending moments caused by column gravity load P and
the lateral drift Δ as lever arm)
2 0 0 6
Bending resistance of moment frame portal under lateral load
-
0
2 Axial resistance of braced frame portal under lateral load
3
9 9
Lateral drift of moment frame is maximum at base
1
erle
4 Lateral drift of braced frame is maximum on top
5 Reduced drift of combined braced/moment frame
c h i
h t G GS
y r i g
Cop

Structure systems
The vertical-lateral framing systems of wall, cantilever, braced frame, and moment
resisting frame, shown from left to right, may be optimized for height and use, including
combinations of systems. The importance to select an efficient system increases with
building height in order to achieve a low weight per floor area ratio for the structure. The
late engineer Fazlur Kahn of Skidmore Owings and Merrill recommended the following
systems for various heights:
Concrete moment resisting frame 20 stories
Steel moment resisting frame 30 stories
Concrete shear wall 35 stories
Braced moment resisting frame 40 stories
Belt truss 60 stories
6
Framed concrete tube 60 stories
Framed steel tube
Braced tube
0 - 2 0 0
80 stories
100 stories
9
Bundled tube 110 stories
1 9
Truss tube without interior columns 120 stories
erle
1 Cellular shear walls
i
2 Exterior shear walls
GS c h3
4
Curved shear walls
Cantilever core with cantilever floors
G
5 Cantilever round core with cantilever floors
t
6 Cantilever core with suspended floors
y r i g h 7
8
Moment resistant frame
Moment frame with two shear cores
Cop
9 Moment frame with single shear core
10 Braced core
11 Braced exterior bays
12 Braced core with outrigger trusses

Tubular systems incorporate lateral resistance into the skin by either some form of
bracing or narrowly spaced columns with moment resisting beam-column connections.
For very tall buildings, bundled tubes increase lateral resistance with interior cell “walls”
to reduce shear lag between exterior skin bracing. The Sears tower in Chicago has a
framed bundled tube structure.
1 Framed tube
2 Braced tube
3 Bundled tube, framed
4 Bundled tube, braced
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Structure systems vs. height
The diagram is based on a study by the late Fazlur Kahn regarding optimal structure
system for buildings of various heights, defined by number of stories.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Structure weights
The amount of structural steel required per floor area is a common measure of efficiency
for steel structures. Comparing various systems demonstrates the importance of
selecting a suitable system. As shown in the diagram, considering gravity load alone, the
structural weight would increase only slightly with height. The effect of lateral load,
however, accelerates the increase dramatically and at a non-linear rate.
1 Structural steel weight related to building height (by Fazlur Kahn)
2 Weight of structural steel per floor area of actual buildings
A Number of stories
B Weight of structural steel in psf (pounds per square foot)
C Weight of structural steel in N/m2
D Weight of structural steel considering floor framing only
6
E Weight of structural steel considering gravity load only
F
G
- 2 0 0
Weight of structural steel for total structure optimized
Weight of structural steel for total structure not optimized
0
H
I
9
Empire State building New York
1
Chrysler building New York
9
erle
J World Trade center New York
c h
K
L
i Sears tower Chicago
Pan Am building New York
GS
M United Nations building New York
G
N US Steel building Pittsburgh
r i g h t O
P
John Hangkock building Chicago
First Interstate building Los Angeles
y
Q Seagram building New York
Cop
R Alcoa building Pittsburgh
S Alcoa building San Francisco
T Bechtel building San Francisco
U Burlington House New York
V IDS Center Minneapolis
W Koenig residence Los Angeles

Floor and roof framing
The layout of horizontal framing is important not only for the transfer of gravity load to
columns but also to resist lateral load. For example, beam framing should transfer
gravity load to columns subject to uplift forces caused by overturn moments. Gravity load
may cancel uplift at least in part to avoid the need for foundation anchorage. It is also
advantageous to design beams to distribute gravity load to girders evenly, rather than all
load to some and none to others.
1 Single-layer system: beams rest on columns
2 Two-layer system: joists rest on beams that rest on columns
3 Three-layer system: joists rest on beams that rest on girders that rest on columns
4 Flush joist and beam
Requires joist connection into side of beam
6
Joists resist rotational buckling of beam
5
Mechanical ducts must run below framing
Layered framing
0 - 2 0 0
9
Provides easy connections
1 9
Main ducts between beam, feeder ducts between joists reduces height
erle
Joists don’t resist rotational buckling of beams
i
A Joist
GS c hB
C
Beam supports joist
Girders supports beam
G
D Column supports beam or girder
t
E Symbol for moment resisting connection common in framing plans
y r i g h
Cop

Beam-column interface
The type of interface between spandrel beam and column on a facade is important
considering architectural and structural implications. Assuming moment resisting
connections, the best structural solution is to frame the beam directly into the column for
effective moment transfer without torsion. If shown on the facade, this expresses most
clearly a moment resisting frame. Beams may run behind the column to express
verticality, or in front of the column to express horizontally; yet both cases generate
torsion in the beam and bending in the column due to eccentricity.
1 Visual expression of frame
2 Axon of beam framed directly into column
3 Section of beam framed directly into column
4 Visual expression of columns
6
5 Axon of beam running behind column
6
7 Visual expression of beam
0 - 2 0 0
Section of beam running behind column yielding a moment couple
9
8 Axon of beam running in front of column
9
1 9
Section of beam running in front of column yielding a moment couple
c h i erle
h t G GS
y r i g
Cop

17
V ERT I CAL SY ST EM S Bending resistant structures include cantilever, moment frame, framed tube, and bundled
tube structures. They resist lateral load by combed axial and bending stresses. Since
bending stress varies from tension to compression with zero stress at the neutral axis,
Be nding Re sist a nt only half the cross section is effectively engaged. This makes them less stiff than shear
2 0 0 6
walls or braced frames, but it provides greater ductility to absorb seismic energy in the
elastic range, much like a flower in the wind. On the other hand, bending resistance
-
0
implies large deformations that may cause damage to non-structural items. Bending
9 9
resistant structures are sometimes combined with other systems, such as braced frames
1
or shear walls, for greater stiffness under moderate load; but moment frames provide
erle
ductility under severe load, after the bracing or shear walls may fail.
c h i
h t G GS
y r i g
Cop
17-1 VERTICAL SYSTEMS Bending Resistant

Cantilever
Cantilever structures consist of towers that rise from the foundation. They resist gravity
load in compression and lateral load in bending and shear, similar to moment frames.
Cantilevers are subject to global bending of the entire tower, whereas moment frames
are subject to localized bending of columns and beams joint by moment resistant joints.
The global bending of cantilever towers increases from minimum on top to maximum at
the base; whereas the local bending of beams and columns in moment frames varies at
each level from positive to negative.
Cantilever towers may be very slender walls, hollow boxes, or solid columns. Compared
to shear walls, which resist lateral load in shear, cantilevers resist primarily in bending.
The most common materials are reinforced concrete and wood poles of pole houses.
6
Floors may also cantilever from the towers. Cantilevers need large foundations to resist
tie the towers together for better stability.
0 - 2 0 0
overturn moments. Cantilever systems of multiple towers may have joint foundations that
1 9 9
Single tower cantilever with cantilever floors
erle
2 Single tower cantilever under lateral load
3 Twin tower cantilever with joined footing for improved stability
c h
4
i Twin tower cantilever under lateral load
GS
5 Single tower cantilever with suspended floors
6 Single tower suspension cantilever under lateral load
i g h t G
y r
Cop

Pirelli tower, Milan (1956-58)
Architect: Ponti, Fornaroli, Rosselli, Valtolina, Dell Orto
Engineer: Arturo Danusso, Pier Luigi Nervi
Facing Milan’s central station across a major urban plaza, the 32-story Pirelli tower rises
prominently above the surrounding cityscape. A central corridor, giving access to offices,
narrows toward both ends in response to reduced traffic. The reinforced concrete
structure features two twin towers in the midsection for lateral resistance in width
direction and triangular tubes at both ends for bilateral resistance. The towers and tubes
also support gravity load. The gravity load of the towers improves their lateral stability
against overturning. The central towers are tapered from top to base, reflecting the
increasing global moment and gravity load. The towers are connected across the central
corridors at each level by strong beams that tie them together for increased stability. In
plan, the central towers are fan-shaped to improve buckling and bending resistance. The
- 2 0 0 6
tubular end towers of triangular plan house exit stairs, service elevators, and ducts.
Concrete rib slabs supported by beams that span between the towers provide column-
0
free office space of 79 and 43 ft (24 and 13 m). The plan and structure give the tower its
9 9
unique appearance, a powerful synergy of form and structure.
1
erle
Floor plan: 18 x 68 m (59 x 223 ft)
i
Height: 127 m (417 ft))
GS c hTypical story height:

Height/width ratio
3.9 m (12.8 feet)
7
i g h t G
y r
Cop

Hypo Bank, Munich (1980)
Architect: Bea and Walter Betz
The design objective for the Hypo Bank headquarters was to create a landmark for
Munich and a unique architectural statement for the bank. Built 1980, the 22 story bank
has 114m height. The structure consist of four tubular concrete towers that support a
platform which supports 15 floors above and 6 floors suspended below. The suspended
floors had been built from top down simultaneous with upper floors being built upwards.
Four towers combined with a platform form a mega-frame to resist gravity and lateral
loads. The four towers include exit stairs in prestressed concrete tubes of 7m diameter
and 50 to 60 cm wall thickness. A fifth tower of 12.5m diameter, houses eight elevators
and mechanical equipment. The support platform consist of prestressed site-cast
concrete of 50cm thick concrete slabs on top and bottom, joined by 1.5m rib walls that
are tied around the towers. The formwork for the platform was assembled on ground and
lifted 45 m by 12 hydraulic jacks.
- 2 0 0 6
The office space consists of three triangular units, joined by a T-shaped center. Two-way
1 9 9 0
beams for office floors are supported by columns above the platform and suspended
below. Three sub-grade levels include parking, security control, and loading stations.
erle
Floor plan: 7 m (23 ft) diameter towers
i
Height: 114 m (374 ft)
GS c hHeight/width ratio
1
16 per tower
Typical upper floor supported by columns above the platform
G
2 Story-high platform forms a mega frame with four towers
r i g h t 3
4
Typical lower floor suspended from the platform
Isometric view of building
Copy 5
6
7
8
Roof plan
Typical office floor framing
Support platform framing
Typical floor plan layout

1 Commerzbank Düsseldorf (1965)
Architect: Paul Schneider-Esleben
This 12-story bank building is located at the boundaries between the old and new
banking district of Düsseldorf, linked by a pedestrian footbridge to an older building
of the bank. The 12-story building above a 2-story podium was initially designed to
allow a drive-in bank at street level. A free-standing service core supports the
pedestrian bridge and makes the link to the office floors. A second stair and
bathroom core is located at the far end of the building, providing undivided and
flexible office space. The curtain wall façade is designed and manufactured using
vehicular technology of insulating sandwich panels. The structure consists of
reinforced concrete. Two rows of square cantilever columns support cantilever
beams and concrete floor slabs. The interior core helps to resist lateral load in
length and width directions, but the exterior core at the other end of the building
resist lateral load in width direction only.
- 2 0 0 6
0
Height:
1 9
Typical story height
44 m (144 ft)
3 m (9.8 ft)
9
erle
Height/width ratio 10 per cantilever
c h2
i Lend Lease House Sydney (1961)
GS
Architect: Harry Seidler
G
This 15-story office tower with north-south orientation of its length axis has
r i g h t movable exterior blinds for sun control. They give the facade an ever-changing
appearance. On sunny mid-days, they are horizontal for optimal sun protection.
Copy On cloudy days, in lowered position, they tend to darken the inside rooms. The
orientation provides inspiring views to the Sydney harbor and a nearby botanical
garden. A two-story showroom with mezzanine floor is located on the ground floor,
above a four-story underground parking garage. The office floors feature
elevators, stair and bathrooms on one end and an exit stair at the opposite end,
providing flexible office floors. Mechanical equipment is in a roof penthouse. The
structure consists of reinforced concrete. Two rows of wall-shape cantilever
columns support cantilever slabs. The cantilever columns resist both gravity and
lateral loads.

Height: 38 m (125 ft)
Height/width ratio 4.7 per twin cantilever

Moment frame
Moment frames consist of one or more portals with columns joint to beams by moment
resistant connections that transmit bending deformation from columns to beam and vice-
versa. Beams and columns act together to resist gravity and lateral loads in synergy and
redundancy. Bending resistance makes moment frames more ductile and flexible than
braced frames or shear walls. The ductile behavior is good to absorb seismic energy, but
increases lateral drift, a challenge for safety and comfort of occupants, and possible
equipment damage.
Moment frames provide optimal planning freedom, with minimal interference of structure.
Office buildings that require adaptable space for changing tenant needs usually use
moment frames. To reduce lateral drift in tall buildings, dual systems may include
bracing or shear walls, usually at an interior core where planning flexibility is not required.
6
Given the high cost of moment-resistant joints, low-rise buildings may provide only some
- 2 0 0
bays with moment resistant frames. The remaining bays, with pin joints only, carry
gravity load and are laterally supported by adjacent moment frames.
0
9 9
Moment frame behavior can be visualized by amplified deformations. The connection of
1
column to beam is usually perpendicular and assumed to remain so after deformation.
erle
Under lateral load, columns with moment joints at both ends assume positive and
i
negative bending at opposite ends, causing S-shapes with inflection points of zero
GS c hbending at mid-span and end rotation that rotates the ends of a connected beam. By
resisting rotation, beams help to resist lateral load. Similarly, a beam subject to bending
G
under gravity load will rotate the columns connected to it and thus engage them in
t
resisting the gravity load. Columns with moment-resistant joints at both ends deform less
y r i g h than columns with only one moment joint. Deformations under gravity and lateral loads
are visualized in the diagrams, with dots showing inflection points of zero bending stress.
Cop
1 portal with hinged joints unable to resist lateral load
2 Moment joints at base, hinge joints at beam, large drift
3 Moment joints at strong beam, hinge joints at base, large drift
4 Moment joints at base and strong beam, drift reduced to half
5 Hinged base, moment joints at beam, beam forms inflection point
6 Gravity load, hinged base, beam moment joints, 2 beam inflection points
7 Lateral load, all moment joints, inflection points at beam and columns
8 Gravity load, all moment joints, inflection points at beam and columns
9 Multi-bay frame deformation under lateral load
10 Multi-bay frame deformation under gravity load
A Inflection point of zero bending stress

Moment-resistant Joints
Moment-resisting joints usually consist of steel or concrete. They join members (usually
column to beam) to transfer bending moments and rotations of one member to the other.
The moment resistant connection makes post and beam act in unison to resist both
gravity and lateral loads. In seismic regions, moment frames must be ductile to absorb
seismic energy without breaking.
Steel moment joints are usually wide-flange beams connected to wide-flange columns.
Generally, post and beam are connected about their strong axis. Semi-rigid joints
connect the strong beam axis to the weak column axis. Moment resistant joints require
stiffener plates welded between column flanges. They resist bending stress of beam
flanges that tend to bend column flanges without stiffener plates. Compact columns with
very thick flanges do not require such stiffener plates. Steel is a ductile material which is
good to absorb seismic energy in the elastic range. Yet the seismic performance of steel
2 0 0 6
joints was challenged by failures during the 1994 Northridge Earthquake. The failure
resulted primarily from joint welds. Research developed solutions for moment-resisting
-
0
steel joints, notably dog-bone beam ends to form plastic hinges to reduce stress at the
joints.
1 9 9
erle
Concrete frames achieve ductile joints by proper steel reinforcing, designed to yield
i
before the concrete crushes in brittle mode. Usually that implies 25% to 50% less steel
GS c hand more concrete than used for balanced design (balanced design has just enough
reinforcing to balance the concrete strength). Ductile design also requires: closely
spaced tie bars near beam/column joints; column rebars to extend through beams; beam
h t G rebars to extend through columns; and column ties to continue through beams.
rig
1 Moment-resisting steel joint at end column
y
2 Moment-resisting steel joint at interior column
C o p 3
4
A
B
Moment-resisting concrete joint at end column
Moment-resisting concrete joint at interior column
Steel wide-flange column
Steel wide-flange beam
C Stiffener plates resist bending stress of beam flanges
D Steel bar welded to column in shop and bolted to beam in field
E Weld, joining beam flange to column
F Steel reinforcing bars in concrete beam
G Steel ties to restrain reinforcing bars from buckling
H Column reinforcing bars to resist compression and bending

Steel framing
Steel framing with wide-flange profiles requires careful orientation of columns in order to
achieve proper strength and stiffness to resist lateral load in both orthogonal directions.
Measured by the moment of inertia, typical wide-flange columns have a stiffness ratio of
about a 3:1 about the x and y-axis, respectively, yet some deep sections have stiffness
ratios up to 50:1, about strong to weak axes. Therefore, column orientation for lateral
resistance is an important design consideration for moment frames. Assuming equal
lateral load and column size, half of the columns should be oriented in either direction.
For unequal loads, column orientations should provide strength proportional to loads.
For example a rectangular building has more wind load on the long than on the short
facade. If wind governs lateral design, this should be considered in column orientation.
Further, column orientation should provide symmetry of stiffness in both directions to
prevent torsion. Torsion would occur for example if one end of a building has columns
6
with greater stiffness than the other end. Also to better resist possible torsion from
- 2 0 0
asymmetric mass distribution, columns should be placed near or at the building edge,
rather than near the center of mass where they have no effective lever arm to resist
0
9
torsion. Column size should also account for setbacks on upper floors, to account for
1 9
asymmetric wind or seismic load resulting from such setbacks.
erle
1 Front view of moment resisting frame with setback floors on top
c h
2
i Column layout in plan for moment resistance in both directions
GS
A Column oriented for lateral support in width direction
B Column oriented for lateral support in length direction
i g h t G
y r
Cop

Casa Terragni, Como, Italy (1936) Commonwealth (formerly Equitable) Building (1944-48)
Architect: Guiseppe Terragni Architect: Pietro Belluschi
With 33.2x33.2x16.6m height, the building is a perfect half cube. The plan is organized The Equitable Building 1948 pioneered the clear expression of a steel
around a central atrium, surrounded by circulation. Terragni used the concrete moment moment frame, a model for many subsequent buildings. With this
frame as organizing grid in a liberal manner, modified as required to meet planning building Beluschi also pioneered the first double glazed aluminum curtain
needs: the 4.75m grid is reduced for circulation and increased for large spaces. Beams wall of simple elegance... The building is a National Historic Landmark of
of variable depth express the respective spans. The front facade is recessed behind a mechanical engineering because it was the first building using heat
veranda to emphasize the frame. Moment frames with shear walls have proven a failsafe pumps for efficient air conditioning... It was the first skyscraper to use
solution in earthquakes prone areas: shear walls provide good stiffness under moderate double-paned glass. The first building with air conditioning completely
load, and the moment frame provides ductility if shear walls fail in sever earthquakes. sealed and the first to use a flush curtain wall design. The first building
completely clad in aluminum. 1982 the American Institute of Architects
awarded the building the 25-year award. The building is a compelling
testimony of Beluschi’s philosophy of simplicity.
.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Seagram building, New York (1954-58)
Architect: Mies van der Rohe, Philip Johnson, Kahn and Jakobs
The 38-story Seagram building is a classic icon of modem architecture. It was the result
of unique cooperation between the client, Samuel Bronford, his daughter, Phyllis Lambert
as planning director, and the architects. The building exemplifies Mies' philosophy of
Baukunst (art and craft of building), with great attention to detail and proportion. The
structure, based on a 28 ft (8.5 m) module, is expressed as colonnade at the base to
signal the entrance. The skin of the mechanical floor on top provides a visual cap. Most
of the structure is concealed behind the curtain wall which eliminates thermal stress and
strain due to outside temperature variations, an important factor in tall structures. The
recessed rear gives the tower its classic proportions of five to three for front and side,
6
respectively. The steel moment frame structure is embedded in concrete for fire
- 2 0 0
protection and added stiffness. The core walls have diagonal bracing up to the 29th floor
for additional wind bracing. Concrete shear walls up to the 17th floor provide additional
0
9
stiffness.
1 9
erle
Floor plan: 84 x 140 feet (26 x 43 m) without extrusion
Height: 525 feet (160m)
c h i
Typical story height: 13.6 feet (4.15m)
GS
Height/width ratio 6.3 without extrusion
G
1 Axon view of tower
r i g h t 2
3
Comer detail of structure and skin
Typical plan with recessed comers to express 3 to 5 proportion
Copy A
B
C
Air conditioning duct as parapet
Glare reducing pink glass appears without color from inside
Bronze cover of steel column embedded in concrete

Crown Zellerbach building, San Francisco (1959)
Architect: SOM and Hertzka and Knowles
Engineer: H. J. Brunnier
The 20-story Crown Zellerbach headquarters building covers about one third of a
triangular site on Market Street, the main street of San Francisco. The building features
a large office wing flanked by an external core for stairs, elevators, bathrooms, and
mechanical ducts. The exterior core gives the office wing a column-free floor area for
optimal space planning flexibility. A planning module of 5.5 feet (1.6 m) provided for
good size office spaces.
The structure is a moment resistant steel frame with wide-flange girders spanning 63 feet
(19 m) across the width of the building, supported by wide-flange columns, spaced 22
feet (6.7 m) on center. Spandrel beams connect the columns in the longitudinal direction.
6
Steel joists, spaced 7 feet (2.1 m), support concrete slabs on cellular metal decks. The
- 2 0 0
joists cantilever at each end of the building. All columns are oriented with their strong
axis to provide moment resistance in the width direction, giving the building much greater
0
9
strength and stiffness in width than in length direction. Since the building is much longer
1 9
than wide, the column orientation is good for wind load which is greater on the long
erle
faced; but it is less effective for seismic load which is greater in length direction. Also
the eccentric service tower causes seismic torsion. The fire exits on both side of the
c h i
service tower are too close together for fire safety and would not be allowed by current
GS
code. The building is supported by a mat foundation, 8 feet (2.4 m) deep, extending the
full width and length of the building. The foundation rests on firm soil 45 feet (13.7 m)
G
below grade under a 2-story parking garage. The steel structure is protected by fire
r i g h t proofing that consists of stucco applied to metal lath wrapped around beams and
columns.
Copy Floor plan:

Height:
Typical story height:
Height/width ratio
201x69 ft (61x21m) without exterior core
320 ft (96m)
13.67 ft (4m)
4.6 without exterior core
A Column spaced 22ft (6.7 m)
B Spandrel beam
C Girder spanning the full width of the building
D Joist spaced 7ft (2.1 m)
E Stiffener plate for moment connection
F Fire proofing on metal lath

Thyssen tower, Düsseldorf (1957-60)
Architect: Hentrich and Petchnigg
Engineer: Kuno Boll
The Thyssen tower’s unique plan of three slabs is a composition with efficient circulation
and good delighting for all offices that are never more than 7m (23 ft) from a window.
The floor area of offices is 62.7% of the gross floor area. Located at the center of town,
the long axis is oriented north-south with a park to the North. The central block includes
the service core and, as tallest block, houses mechanical and elevator equipment in the
2 0 0 6
top floors of this 25-story tower. Parking for 280 cars is in the underground garage,
rapped around the building. The long facades feature glass curtain walls; the narrow end
-
0
facades are glad in stainless steel. The steel frame structure is embedded in concrete
9 9
for fire protection and to provide additional stiffness. The columns consist of steel pipes
1
produced by the building owner. The structural module is 7x4.2 m (23x14 ft), with some
erle
variation between central and outer slabs. Braced end walls provide some additional
i
stiffness to resist wind load on the long building sides. The exterior composition of the
GS c hbuilding, expressing the internal organization, has earned the nickname “Drei-Scheiben
Haus” (Three-slab-house). The pristine design, combining American know-how with
European sophistication stands as an icon of the modern movement in Europe.
h t G
rig
Floor plan: 21x80 m (70x226 feet)
y
Height: 94 m (308 feet)
C o p Typical story height:

Height/width ratio
12.2 feet (3.7 m)
4.48

Framed Tube
Framed tubes are a variation of moment frames, wrapping the building with a “wall” of
closely spaced columns and short spandrel beams. To place the lateral resistance
system on the façade rather then at the interior gives it a broader base for greater
stability as well as improved rotational resistance. In addition, the lateral resisting system
on the façade allows smaller columns on the interior to carry gravity load only. Further,
designing floors and roof to span the full width of a building can make the interior
completely column free for optimal flexibility. A major challenge of framed tubes is the
high cost of numerous moment resistant joints between closely spaced columns and
beams. To minimize this adverse cost factor, designers often use prefab methods to
weld the joints in the fabrication hop rather than on the job site. This process also
improves quality control and reliability.
6
1 Framed tube without interior core
2
3
Framed tube with interior core
Global stress diagram of framed tube
0 - 2 0 0
9
4 Framed tube with belt and top truss for additional stiffness
5
1 9
Prefab frame with joints located at beam inflection point of zero bending
erle
6 Prefab element ready for assembly
i
A Reduced shear resistance (shear lag) at hollow interior
GS c hB
C
Peak axial force from overturn moment
Pin joint at inflection point of zero beam bending stress
i g h t G
y r
Cop

CBS Tower New York (1961-650
Architect: Eero Saarinnen
Engineer: Paul Weidlinger
The 38-story CBS tower is a stark vertical extrusion of the rectangular floor plan.
Columns forming a framed tube are expressed as triangular extrusions on the upper
floors and diamond shaped on the ground floor. The triangular columns include niches
for mechanical ducts and pipes. The niches decrease from top to bottom with the
decreasing duct sizes that run down from the mechanical room on the top floor. The
decreasing niches result in increasing net column size that coincides with increasing load
as it accumulates from top down. Concrete floors span between the walls of a central
core and the framed tube, providing a column-free donut-shape floor space for flexible
- 2 0 0 6
use. The four sides facing the core feature one-way rib slabs, but the four corners have
two-way waffle slabs, designed to make the transition from one direction to the other.
0
Glad in black granite the closely spaced triangular columns express a stark verticality,
9 9
perforated with regular windows on all but the top and ground floors. The top mechanical
1
floor has ventilation louvers instead of windows; the ground floors have taller windows
erle
and doors. The articulation of top and bottom of the façade emphasizes the most
i
prominent part of the building, a strategy often use for the design of tall buildings.
GS c hFloor plan:
Height:
155x125 feet (47x38m)
494 feet (151m)
G
Typical story height: 12 feet (3.66m)
r i g h t Floor-to-ceiling height:
Height/width ratio
8.75 feet (2.67m)
3.9
Copy A
B
C
Column profile at top floor
Column profile at lower floors
Column profile at ground floor

World Trade Center, New York (1977
(demolished by terrorists 9-11-2001)
Architect: Minoru Yamasaki and E. Roth
Engineer: Skilling, Helle, Jackson, Robertson
The World Trade center housed 50,000 employees and up to 80,000 visitors daily in two
110-story towers. Both towers, in diagonal juxtaposition, were vertical extrusions of
square plans, with very closely spaced steel columns. Each tower had two-story
mechanical spaces on top, near the bottom, and two distributed at 1/3 intervals, with
elevator sky-lobbies two floors above each. Each tower had 100 passenger and four
service elevators. Each sky-lobby was reached by 11 or 12 elevators from ground floor;
with five express elevators non-stop to the 107th and 110th floors. Since elevators are
stacked, 56 shafts needed, take 13% floor area on each floor. The framed tube structure
- 2 0 0 6
consisted of 56 box steel columns on each façade, joint at each floor by spandrel beams
with moment resistant connections. This giant Vierendeel frame was assembled from
0
prefab elements of three two-story columns with beam and column joints at mid-span and
9 9
mid-height where inflection points of zero bending occur under lateral load. Combined
1
with rigid floor diaphragms, the towers formed torsion-resistant framed tubes that
erle
cantilever from a five-story underground structure that houses train and subway stations
i
as well as parking for 2,000 cars. Although the framed tube columns overall dimensions
GS c hare constant, their wall thickness increases from top to bottom in response to increasing
loads. Floor truss joists span from the framed tube to columns around the central core.
Mechanical ducts run between truss joists for reduced story height. The core columns
i g h t G are designed to carry gravity load only. The framed tube resisted both lateral and gravity
load.
y r
Cop
Floor plan (square): 208x208 feet (63.4x63.4 m)
Height: 1361 feet (415m)
Floor-to-ceiling height 8.6 feet (2.62m)
Height/width ratio 6.5
1 Axon of tower
2 Typical floor framing plan
3 Typical prefab two-story façade assembly
4 Typical framed tube column size and spacing

Bundled Tube
Bundled tube structures are composed of tubes framed by closely spaced columns joined
to beams to form moment frames. The bundled tubes resulting from the rows of columns
add lateral resistance to the structure, transferring shear between exterior columns
subject to tension and compression under lateral load. This shear transfer makes it
possible for the exterior columns to act in synergetic unison, whereas independent
columns would act alone to provide much less lateral resistance. Bundled tubes transfer
shear not only through exterior frame “walls” but also through interior cell “walls” thereby
reducing shear lag.
An alternative to framed bundled tube are braced bundled tube systems. However,
though they provide greater stiffness, the braces disrupt spatial flow between interior
columns. Regarding plan geometry, bundled tubes may have bundles of square,
6
rectangular, or triangular polygons that are repeatable. However, hexagonal polygons
would be less efficient
1 Square tube modules
0 - 2 0 0
2
3
9
Triangular tube modules
1 9
Hexagonal tubes would be less effective to reduce shear lag
erle
4 Framed tube shear lag
i
5 Bundled tube with reduced shear lag
GS c hA
B
Shear lag between connecting shear walls
Peak resistance at shear wall
i g h t G
y r
Cop

Sears tower, Chicago (1973)
Architect/ Engineer: Skidmore, Owings and Merrill
With 110 stories, the Sears Tower was the tallest building in the world for many years
and occupies an entire city block on the southwest of Chicago’s loop. The tower starts at
ground level with nine square modules of 75 feet (22.9m) each. The nine modules
gradually reduce to a twin module on top in response to needed office space and also to
reduce wind resistance and overturn moments. The large areas of the lower floors are
occupied by Sears; the smaller floors at higher levels serve smaller rental needs.
Elevators serve the building in three zones of 30 to 40 stories separated by sky-lobbies
that are reached by double deck elevators express elevators. The building façade is glad
in black aluminum and tinted glass. The structure is anchored to a five-story
underground structure. Nine bundled tubes are separated by rows of columns, spaced
15 feet (4.6m) on center. The columns, welded to beams, form moment resisting portals
- 2 0 0 6
to transfer global shear under lateral load from compressed to tensed side of the
structure, to reduce lateral drift. This shear transfer between exterior walls reduces
0
“shear lag” and gives the bundled tube greater strength and stiffness to resist lateral
9 9
loads. The bundled tube concept conceived facilitates the setbacks as the floors get
1
smaller toward the top. Belt trusses at three levels in conjunction with mechanical floors
erle
reduce lateral deflection by about 15 percent and help distribute uneven gravity load
i
caused by floor setbacks. The horizontal floor framing consists of trusses that span 75
GS c hfeet (23m) between columns and support concrete slabs on metal deck. The one-way
floor trusses of 40 inch (1m) depth change direction every 6th floor to redistribute the
gravity load to all columns. Trusses consist of top and bottom T-bars, connected by twin
i g h t G angle web bars. They allow mechanical ducts between top and bottom chords. The
small truss depth was possible, using composite action; shear studs engage the concrete
y r slab in compression for increased resistance.
Cop
Floor plan at ground: 225 x 225 feet (69 x 69m)
Height: 1,450 feet (442m)
Height/width ratio 6.4
1 Tower axon
2 Base floor plan
3 Floor framing
4 Stress diagram of single framed tube with shear lag between walls
5 Stress diagram of bundled tube with reduced shear lag
6 Floor plan at ground floor
7 Floor plan starting at 51st floor
8 Floor plan starting at 66th floor
9 Floor plan of top floors

19
V e rt ic a l Syst e m s Vertical systems, suspended, also referred to as suspended high-rise structures, are
different from suspension structures like suspension bridges, which are draped from two
suspension points; suspended high-rise structures hang usually about vertically from top.
Suspe nde d
6
A rational for suspended high-rise structures is to free the ground floor from obstructions.
0 - 2 0 0
Other architectural and structural reasons are described on the next page.
9
Regarding Lateral load, the challenge of suspended high-rise is usually a narrow footprint
9
and slender aspect ratio. Thus their behavior is comparable to a tree, where the drunk
1
erle
resists load primarily in bending and large roots are required to resist overturning.
Properly designed, the narrow aspect ratio can enhance ductility to make the structure
c h i
behave like a flower in the wind to reduce seismic forces.
h t G GS
y r i g
Cop
19-1 VERTICAL SYSTEMS Suspended

Suspension rational
At first glance suspended high-rise structures seem irrational, given the load-path detour:
gravity load travels to the top and then down to the foundation. However, as described
below, there are advantages, both architectural and structural, that justify this detour.
Understanding the pros and cons and their careful evaluation are essential for design.
Challenges
• Load path detour: load travels up to the top, then down to foundation
• Combined hanger / column deflection yields large differential deflection
Architectural rational
• Less columns at ground floor provides planning flexibility and unobstructed view
6
• Facilitates top down future expansion with less operation interference
0 - 2 0 0
• Small hangers instead of large columns improve flexibility and view
9
Structural rational
1 9
• Eliminates buckling in hangers, replacing compression with tension
erle
• High-strength hangers replace large compression columns
i
• Floors may be built on ground and raised after completion
GS c h• Concentration of compression to a few large columns minimizes buckling
G
Design options
r i g h t • Multiple towers with joint footing to improve overturning resistance

• Multiple stacks to limit differential deflection
y
• Adjust hangers for DL and partial LL to reduce deflection
Cop
• Prestress hangers to reduce deflection to half
1 Gravity load path

Load travels to top, then down to foundation
2 Differential deflection is cumulative
Shortening of columns and elongation of hangers are additive
3 Prestress can reduce deflection to half
Top resists half the load through increase of prestress
Bottom resists half the load through decrease of prestress
4 Ground anchors for improved stability
(assuming hangers as ground anchors are ok)

Design options
Suspended high-rise structures may be designed in various configurations with distinct
limitations and implications regarding behavior. The following description provides
guidelines for rational design, starting with the introduction of a terminology, followed by
implications of various design options:
• Single towers (one vertical support)
• Multiple towers (several vertical supports)
• Single stacks (one set of floors)
• Multiple stacks (several sets of floors)
The effect of these options are described and illustrated as follows:
6
1 Single tower / single stack
0
Single towers require large footing like a tree to resist overturning
2 Multi towers
9 0 - 2 0
9
Multiple towers with joint footing increase stability
erle 1
Twin stacks
c h i Twin stacks reduce the length of hangers and thus differential deflection
GS
(ten stories per stack limits differential deflection to < 2 inch (50 mm)
G
4 Twin stacks / towers
r i g h t Twin stacks reduce the length of hangers and thus differential deflection
Twin towers with joint footing increase stability
Copy 5
6
Triple stacks
Three or more stacks limit hanger length and thus differential deflection
Triple stacks / twin towers

Three or more stacks limit hanger length and thus differential deflection
Two or more towers with joint footing increase stability
Limits
An important limit for suspended high-rise structures is the limited number of floors per
stack. More than ten floors per stack would cause unacceptable differential deflections.
Conventional columns in compression are subject to about equal strain under load.
Suspended high-rise structures are subject to greater differential deflection since hangers
elongate but columns shorten under gravity load. Without buckling, the high tensile
stress of hangers causes greater strain which further increases differential deflection.

Case studies
Westcoast Transmission Tower, Vancouver (1969)
Architect: Rhone and Iredale
Engineer: Bogue Babicki
The 12-story tower, initially designed and built as Westcoast Transmission headquarters,
has become an architectural icon of Vancouver. With support of the City of Vancouver,
the historically significant building was converted in 2005 to 180 unique residential suites
in studio, one and two bedroom configurations. The suspension concept was selected to
provide an unobstructed view to the beautiful bay of Vancouver. According to the Bogue
6
Babicki, the suspension option was also more economical than a conventional alternative
- 2 0 0
they had considered. The suspended structure, stating 30 feet (9 m) above grade
provided unobstructed views at ground level to the beautiful bay of Vancouver. The
0
9
tower is supported by a site-cast concrete core, 36 feet (11 m) square. The floors are
1 9
suspended by 12 cables. Each cable consists of two 2 7/8” (73 mm) diameter strands.
erle
The sloping guy cables have two additional 2 ½” (64 mm) diameter strands (the 45
degree slope increased their vector force by 1.414).
c h i
GS
Size: 108x108 feet (33 x33 m)
Core size: 36x 36 feet (11x11 m))
G
Height 12 stories, 224’ (68 m)
r i g h t Typical story height:

Core height/width ratio:
12 feet (3.65 m)
6.2
Copy 1
2
3
Section
Exploded axon
Floor framing

BMW Headquarters Munich (1972)
Architect: Karl Schwanzer
Engineer: Helmut Bomhard
The Viennese architect Karl Schwanzer won the international design competition for the
BMW tower with his idea to represent the automobile company in form of a four-cylinder
engine. Four cylinders are suspended from an assembly of four semi-cylindrical concrete
cores by means of hangers, suspended from concrete cores of stairs, elevators, etc. The
core extends as four cylinders on top of the floor stacks. Each floor is supported by a
hanger at its center and stabilized by the core. To keep differential deflection within
acceptable limits, the tower is partitioned into two stacks of eleven and seven office floors
of the lower and upper stacks, respectively. Eight elevators, stairs and services are
located in the core. Except for the four central hangers, the office space around the core
- 2 0 0 6
is free from columns to provide highly flexible office areas. Construction of the tower
started with the central core in conventional method; but then proceeded from top down.
0
Post-tensioned concrete floor plates, cast on the ground, where lifted up by hydraulic
9 9
means; starting with the top floor, followed by successive floors downward. Silver
1
gladding exterior conveys a sophisticated high-tech image, true to the BMW philosophy.
c h i
Size:
erle
Core size
52, 30 m (172 feet) diameter
24.4 m (80 feet)
GS
Height: 18 suspended stories, 101 m (331 feet)
Typical Story height: 3.82 m (10.8 feet)
i g h t G Core height/width ratio: 4.1
y r
Cop

Standard Bank Center, Johannesburg (1968)
Architect: Hentrich and Petschnigg
Engineer: Ove Arup and Partners
The Standard Bank Center is located in the financial center of Johannesburg. Given the
dense surroundings, the design objective was to access the center via an open plaza
with the least amount of bulk and obstructions. The response to this objective was a
suspended structure. The central support core only keeps the plaza level open for free
and spacious access. The suspension system also facilitated construction at the dense
urban surrounding. After the central core was built, floors were suspended from three
cantilevers. To limit differential deflection, the building is organized into three stacks of
6
nine office floors each, suspended from concrete cantilever beams of 18 feet (5.4 m)
- 2 0 0
depth. The cantilever beams are attached to the outside face of the concrete core by
shear connection. The cantilever floors house the mechanical equipment and
0
9
transformer stations. Basement floors for computer rooms and parking provide stability
9
for the central core.
1
erle
Floor size: 112x112 feet (34.29x34.28 m)
c h i
Core size: 48x48 feet (14.63x14.63 m)
GS
Building height: 27 stories, 456 feet (139 m)
Core height: 520 feet (158.5 m)
G
Core height/width ratio: 10.8
r i g h t
Copy

Hon Kong and Shanghai Bank (1985)
Architect: Norman Foster
Engineer: Ove Arup
The design of the Hong Kong and Shanghai Bank emerged from a competition among
seven invited architects. Foster's winning scheme is a suspension system intended to
provide large public space at ground level without interior columns. The large floor area
of 55x70m (180'x230') is supported by 8 Vierendeel towers, each consisting of four round
columns spaced 5.1x4.8 m (17'x16’) and connected at each level with tapered beams.
The floors are suspended from twin suspension trusses which span the towers and
cantilever from them on both sides to support service modules and exit stairs. A large
floor area of 33.6x55 m (110'x180') between the towers are disrupted by only eight
hangers, an additional benefit of the suspension scheme, besides the open ground floor.
2 0 0 6
The space between two-story high suspension trusses serves as focal point of each
stack of floors, as reception, conference and dining areas and lead to open recreation
-
0
terraces with dramatic views of Hong Kong.
9 9
The maximum mast pipe diameter is 1400 mm (55”0 and 100 mm (3.9”) thick
1
The maximum hanger pipe diameter is 400 mm (16”) and 60 mm (2.4”) thick
c h i erle
h t G GS
y r i g
Cop

The suspension trusses and X-bracing perpendicular to them are also intended as belt
trusses to reduce drift under lateral load. However, since the Vierendeel towers are
moment resistant, the belt trusses are less effective than they would in conjunction with
truss towers.
Size: 55x70m (180'x230')

Tower axis distance 38 m (126’)
Height 35 floors, 180 m (590’)
Typical story height: 3.9 m (12.8’)
Height/width ratio: 4.7
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Federal Reserve Bank, Minneapolis (1971-73)
Architect: Gunnar Birkerts
Engineer: Skilling, Helle, Christiansen, Robertson
The Federal Reserve Bank features a structure similar to suspension bridges. The floors
are suspended from parabolic “cables”. However, the “cables” are actually wide-flange
steel sections of parabolic curvature to balance the distributed floor loads. A major
6
reason to suspend the building from two towers was to keep the bank vaults located
- 2 0 0
below grade free of columns. Wide flange parabolic suspenders of 37 inch (94eemm)
span 328 feet (100m) between two concrete towers. Trusses on top of the towers resist
0
9
the lateral trust of the parabolic suspenders. Floors above the suspenders are supported
9
by compression columns, whereas those below are suspended by tension hangers. The
1
erle
façade treatment reflects the compressive and tensile support zones by different recess
of the glass line with respect to curtain wall mullions. Floor construction of concrete slabs
c h i
rests on steel trusses that span the 60 feet (18m) width without interior columns.
GS
Size: 335x60 feet (102x18 m)
G
Span between towers 275 feet (84 m)
r i g h t Height
Typical story height
220 feet (67 m)
12.5 feet (3.8 m)
Copy Height/width ratio: 3.7

23 Material
Concrete is a versatile material that can be molded into many forms. It was first known
in ancient Rome. Quarrying limestone for mortar, the Romans discovered that burning
the limestone mixed with silica and alumina yields cement stronger and more adhesive
than ordinary lime mortar. The new material also could be used for underwater
construction. Mixing the cement with sand and other materials, the Roman's invented
the first concrete and used it widely in their construction, often filled between masonry.
Conc re t e The technology of concrete construction was lost with the fall of the Roman Empire. Only
toward the end of the eighteenth century did British inventors experiment to develop
concrete again. In 1825, Joseph Aspdin patented Portland cement which he named after
6
Portland limestone of similar color. The material was soon in wide use and the name
- 2 0 0
Portland cement is still common today. It consists of lime, silica, and alumina, burned to
clinkers in a furnace at about 3000º F (1650º C), and then crushed to a fine powder.
0
9 9
Concrete, consisting of cement, sand, and gravel mixed with water, is strong in
1
compression, but very weak in tension and shear. Thus, concrete by itself is limited to
erle
applications subject to compressive stress only. This limit was soon recognized and by
i
1850 several inventors experimented with adding reinforcing steel to concrete. In 1867
GS c hthe French gardener, Joseph Monier, obtained a patent for flower pots made of
reinforced concrete. He went on to build water tanks and even bridges of reinforced
G
concrete. Monier is credited to invent reinforced concrete...
r ig h t As any material, concrete has advantages and disadvantages. Concrete ingredients are
widely available and rather inexpensive. Concrete combines high compressive strength
Co p y with good corrosion and abrasion resistance. It is incombustible and can be molded in
many forms and shapes. Concrete's main disadvantage is its weakness in resisting
tension and shear. Steel reinforcing needed to absorb tensile stress can be expensive.
Concrete has no form by itself and requires formwork that also adds much to its cost.
The heavy weight of concrete yields high seismic forces but is good to resist wind uplift.
Concrete is inherently brittle with little capacity to dissipate seismic energy. However,
concrete frames with ductile reinforcing can dissipate seismic energy. The inherent fire
resistance of concrete is an obvious advantage in some applications.
Today, concrete serves many applications usually with reinforcing. In buildings, concrete
is used for items like footings and retaining walls, paving, walls, floors, and roofs.
Concrete is also used for moment resistant frames, arches, folded plates and shells.
Apart from buildings, many civil engineering structures such as dams, bridges, highways,
tunnels, and power plants are of concrete.
23-1 MATERIAL Concrete

Concrete properties
Normal concrete has compressive strengths of 2 to 6 ksi (14 to 41 MPa) and high
strength concrete up to 19 ksi (131 MPa). Low-strength concrete is used for foundations.
Concrete strength is determined by the water-cement ratio (usually 0.6) and the cement-
sand-gravel ratio (usually 1-2-3). Specified compressive strength of concrete ƒ‘c, usually
reached after 28 days, defines concrete strength. By the strength method (ultimate
strength method) a structure is designed to 85% of the specified compressive strength
ƒ‘c, with factored loads as safety factor. By the working stress method, a structure is
designed to allowable stress, i.e., a fraction of the specified compressive strength ƒ‘c.
Allowable concrete stress for working stress method

Compressive bending stress 0.45 ƒ‘c
Bearing stress (full area): 0.25 ƒ‘c
Bearing stress (1/3 area):
Shear stress without reinforcing:
- 2 0 0
beam
6 0.375 ƒ‘c
1.1 ƒ‘c1/2
0
joist 1.2 ƒ‘c1/2
1 9
Shear stress, with reinforcing:
9 footing and slab 2.0 ƒ‘c1/2
5 ƒ‘c1/2
erle
Elastic modulus (w = concrete density in pcf): w 33 ƒ‘c1/2
1.5
c h i
Temperature increase causes expansion of concrete defined by the thermal coefficient
GS
α=5.5x10-6 in/in/°F (3.1x10-6 m/m/°C). Hence, concrete slabs need temperature
reinforcing to prevent cracks due to uneven expansion. Concrete also has creep
i g h t G deformation over time, mostly during the first year. Concrete shrinks about 1.3% due to
loss of moisture, notably during curing. The temperature reinforcing helps to reduce
r
shrinkage cracks as well. Density of concrete is determined by the type of aggregate.
Copy Light-weight concrete weighs about 100 pcf (1602 kg/m3). Normal concrete 145 pcf
(2323 kg/m3) without reinforcing and 150 pcf (2403 kg/m3) with reinforcing.
Concrete has good fire resistance if reinforcing steel is covered sufficiently.
An 8 in (20 cm) wall provides 4 hours and a 4 in (10 cm) wall 2 hours fire resistance.
1 Stress-strain curves for concrete
2 Concrete creep (deflection with time)
3 Concrete strength increase with time as percentage of 28-day strength
A Point defining line of E-module on curve
B Elastic limit of idealized line for working stress method
C Idealized line for strength method
D Actual stress-strain curve
E Elastic modulus, defined as the slope from 0 to 0.5 ƒ‘c
ε Unit strain, in/in (m/m)
F Unit creep strain, in/in (m/m)
G Days after pouring concrete

Cement comes in bags of 1 ft3 (.028 m3), classified by ASTM-C150 as:
Type I Normal cement (for most general concrete)
Type II Moderate resistance to sulfate attack
Type III High early strength
Type IV Low heat (minimizes heat in mass concrete, like dams)
Type V High resistance to sulfate attack
Types IA, IIA, IIIA correspond to I, II, III, but include air-entraining additives for improved
workability and frost resistance.
Water must be clean, free of organic material, alkali, oil, and sulfate. The water-cement
ratio defines the strength and workability of concrete. Low water content yields high
strength, but is difficult to work. Typical water ratios are 0.4 to 0.6, verified by a slump
test. For this test, a metal cone is filled with concrete and tamped. Lifting the cone
6
slumps the concrete to under 3 in (7 cm) for foundations and walls, and 4 in (10 cm) for
columns and beams.
- 2 0 0
Aggregate should be clean and free of organic material. Fine aggregate (sand) is less
0
9
than 1/4 in (6 mm). Coarse aggregate (gravel or crushed rock) is used in normal
1 9
concrete. Lightweight concrete has aggregate of shale, slate, or slag. Perlite and
erle
Vermiculite are aggregates for insulating concrete.
i
Admixtures are substances added to concrete to modify its properties:
GS c h•
•
Air entrained agents improve workability and frost resistance
Accelerators reduce the curing time and increase early strength
G
• Retarders slow the curing and allow more time to work the concrete
r i g h t •
•
Plasticizers improve the workability of concrete
Colors and pigments add colors to concrete
Copy Curing of concrete is a process of hydration until it reaches its full strength. Although
this process may take several months, the design strength is reached after 28 days.
During the curing process the concrete should remain moist. Premature drying results in
reduced strength. Exposed concrete surfaces should be repeatedly sprayed with water
or covered with a protective membrane during curing. This is most important in hot or
windy climates. The curing process accelerates in hot temperatures and slows down in
cold temperatures. Concrete shrinks about 2 % during curing. This may cause cracks.
Synthetic fibers of 1/8 to 3/4 in (3 to 20 mm) are increasingly added to improve tensile
strength and reduce cracking of concrete.
1 Concrete compressive strength defined by water-cement ratio
2 Slump test: sheet metal cone and slumped concrete, C = slump
3 Maximum aggregate sizes: 1/3 of slab, 1/5 of wall, 3/4 of bar spacing
A Compressive strength of normal concrete
B Compressive strength of air entrained concrete
C Slump is the amount the wet concrete settles

Reinforced concrete
Concrete is strong in compression, but weak in tension and, when cracked, has zero
tensile strength. Under tensile stress, concrete requires reinforcement with deformed
bars or welded wire fabric. Concrete and steel are compatible, with thermal coefficients
of α = 6x10-6/F° and 6.5x10-6/F° for concrete and steel, respectively. With different
thermal expansions, major thermal stress would result. Some temperature reinforcement
is required to prevent cracks. Concrete protects the embedded steel from fire and
corrosion, but cracks cause steel corrosion by exposing it to humidity.
Deformed bars have round cross sections with ribs to bond with concrete. Under certain
conditions bars need a hook at the end to resist slippage. Bar sizes are designated by
numbers 3 to 18. Up to size 8, bar numbers correspond to the bar diameter in eighth of
an inch (No. 7 = 7/8 in. Bars are available in the following grades and corresponding yield
6
strengths ƒy
Grade
ƒy ksi (MPa)
40
40 (275)
50
50
0 - 2
(345)
0 0 60
60 (414)
90∗
90∗ (620∗)
9
∗ Available for bars No. 14 and 18 only (for compression reinforcement).
1 9
erle
Properties of reinforcing bars
Diameter Area Weight
c h i
Size in mm in2 mm2 lb. /ft kg/m
GS
3 0.375 9.50 0.11 71 0.376 0.560
4 0.500 12.70 0.20 129 0.668 0.994
G
5 0.625 15.88 0.31 200 1.043 1.552
r i g h t 6
7
0.750
0.875
19.05
22.22
0.44
0.60
284
387
1.502
2.044
2.235
3.042
Copy 8
9
10
11
14
1.00
1.128
1.270
1.410
1.693
25.40
28.65
32.26
35.81
43.00
0.79
1.00
1.27
1.56
2.25
510
6.45
819
1006
1452
2.670
3.400
4.303
5.313
7.650
3.973
5.060
6.404
7.907
11.380
18 2.257 57.33 4.00 2581 13.600 20.240
1 Deformed bars with stamp for mill, bar #, steel type, and grade
3 Bar hooks of 90°, 180°, and 135°; the latter for stirrups and ties only
2 Minimum bar bend defined by bar diameter
A Hook length: 6D, stirrups and ties; 12D all others; or min. 2.5 in (6 cm)
B Hook length: 4D; or min. 2.5 in (6 cm)
C Hook length: 6D (10D for seismic regions); or min. 2.5 in (6 cm)
D Bar diameter
E Bend diameter: 4D, No. 5 bars and smaller for stirrups and ties only;
Other bars: 6D, No. 3 to 8; 8.5D, No. 9 to 11; 10.5D, No.14 and 18

Bar cover is the distance between bar edges and the outside surface of concrete.
Structurally, reinforcing is most effective near the surface to resist cracking and bending.
The distance from the neutral axis near the center increases the resisting lever arm for
steel. This makes it more effective to resist bending. However, bars placed too close to
the surface are more susceptible to corrosion and are poorly protected against fire. The
ACI code defines the minimum cover for various members and exposure conditions for
the purpose of fire and corrosion resistance. This is particularly important for members in
contact with soil, such as foundations, or basement walls.
Bar spacing must be wide enough to allow wet concrete to flow freely and to transfer
stress between spliced bars. But spacing should be close enough to provide effective
reinforcing. The ACI code defines upper and lower limits for bar spacing.
6
The diagrams show minimum bar cover and spacing for typical concrete structures,
1 Beam
0 - 0
including beam, post, foundation, and slab and plate.
2 0
2
3
9
Slab or plate
1
Column
9
erle
4 Foundation
i
5 Wall
GS c hA
B
Minimum bar covers: 1.5” for beam and post; 1” for joist
Minimum bar spacing for beam: 3/4” or 1.33 max. bar φ
i g h t G C Minimum bar cover for slab and plate: 3/4” for #5 bar and smaller
(1.5” when exposed to weather); 2” for #6 bars and larger
r
D Minimum bar cover for foundation: 3”
Copy Welded Wire Fabric is common as reinforcement for slabs on grade and thin slabs. It
consists of orthogonal welded wire mesh. Wires are smooth or deformed for better
bonding and come in yield strengths from 56 to 70 ksi (386 to 483 MPa). The largest
wire has 0.2 sq in area and 1/2 in (13 mm) diameter. A typical welded wire fabric
designation is 4x6-W10xW20, implying:
4x6 Wire spacing (in)
W10xW20 wire size and type (W for smooth, D for deformed wires)

Beam Reinforcement
Concrete beams require reinforcement for bending and shear in correlation with the
respective stress patterns. This is illustrated for a simply supported beam under uniform
load and for other beams on the next page.
Bending reinforcement is placed where the bending moment causes tensile stress. A
simply supported beam under uniform gravity load deforms downward to generate
compression on top and tension at the bottom. Thus, bending reinforcement is placed at
the bottom. Beams with negative bending require tensile reinforcement on top. This is
the case in beams with moment resistant supports, cantilever beams, and beams
continuing over three or more supports. Some beams may require additional bars at
mid-span or over supports to resist increased bending moment. Beams of limited depth
also require compressive reinforcement to make up for insufficient concrete. Some
- 2 0 0 6
reinforcement bars have hooks at both ends to anchor them to the concrete if the bond
length between steel and concrete is insufficient. Deformed bars usually don’t need
0
hooks, given sufficient bond length. Temperature reinforcement resists stress caused by
9 9
temperature variation and shrinkage during curing.
1
erle
Shear reinforcement is placed where the shear stress exceeds the shear strength of
concrete which is very small compared to compressive strength. Beams under uniform
c h i
gravity load have maximum shear at supports which decreases to zero at mid-span.
GS
Thus, shear reinforcement in form of stirrups is closely space near the supports and
spacing increases toward mid-span. Stirrups are usually vertical for convenience,
G
though combined horizontal and vertical shear stresses generate diagonal tension which
r i g h t may cause diagonal cracks near the supports. Small longitudinal bars on top of a beam
tie the stirrups together.
Copy 1
2
3
4
Shear diagram: maximum shear at supports and zero at mid-span
Isostatic or principal stress lines: diagonal tension, dotted, near support
Side view of beam with reinforcement
Axon view of beam with reinforcement
A Bottom steel bars resist tensile stress
B Stirrups resist shear stress which, for uniform load, is maximum at the supports
and zero at mid-span

Wall
The unsupported height to width ratio of bearing walls should not exceed 25 with 6 in (15
cm) minimum thickness - 8 in (20 cm) for basement walls. Non-bearing walls may be 4 in
(10 cm) thick. Walls of 10 in (25 cm) or thicker should have 2 layers of reinforcement.
Unless reinforcement is determined to be greater for a given condition, the following
minimum reinforcement shall be provided as a percentage of the wall cross-section area.
Horizontal reinforcement: 0.25% min.
Vertical reinforcement: 0.15% min.
Additional reinforcement is required at wall tops, corners, around all openings, as well as
foundations.
Dowel bars connect walls to foundations, floor and roof slabs. They should overlap with
rebars at least 40 bar diameters or the n=minimum computed bond length.
1
2
Exterior wall with flat slab
Interior wall with flat slab
- 2 0 0 6
3
4
1 9 9
Exterior wall with rib or waffle slab
Interior wall with rib or waffle slab
0
erle
5 Exterior wall with foundation and slab on grade
6 Interior wall with foundation and slab on grade
S c h
A
B i Wall with 2 layers of reinforcement
Flat slab with dowel bars connected to wall
h t G G C
D
Rib or waffle slab with dowel bars connected to wall
Slab on grade with construction joint at wall
i g
E Gravel bed under slab on grade
p y r F Foundation with dowel bars and key
Co

Slab
Depending on the support conditions, concrete slabs may span one-way or two-ways. If
supports are on two opposite sides, a one-way slab is the only option. For a slab
supported along all edges and of approximately equal span in both directions, a two-way
slab is preferred and more efficient. However, if spans in the two directions are different,
a one-way slab is better since deflection increases with the fourth power of span, causing
16 times greater deflection for double spans. In two-way slabs of unequal spans, the
rebars spanning the short direction carry most of the load and bars spanning the long
direction are ineffective. The ratio between short and long span should not exceed 1:2,
but is most effective at 1:1.
The span capacity for slabs is about 20 or 30 feet for one-way and two-way slabs,
respectively. Slabs exceeding those limits require intermediate beams or joists. One-
1 One-way slab supported by two edge beams
- 2 0 6
way and two-way slabs are shown on the left and right, respectively.
0
2
3
1 9 9 0
Two-way slab supported by four edge beams
One-way beams (supporting slab) supported by two edge beams
erle
4 Two-way beams (supporting slab) supported by four edge beams
5 One-way Rib-slab (pan joist) supported by beams
c h i Slab depth 2.5” to 4.5” (6 to 10 cm)
GS
Total depth 10” to 24” (25 to 60 cm); L/d= 20-28
Rib width 5” to 10” (13 to 25 cm)
G
Rib spacing 2’ to 3’ (0.6 to 1 m)
r i g h t 6 Two-way waffle slab with solid panels over columns to resist shear:
Slab depth 2.5” to 4.5” (6 to 10 cm)
Copy Total depth 10” to 24” (25 to 60 cm); L/d = 33

Waffle size 2’ to 5’ (0.6 to 1.5 m)

Slab and plate
Depending on span and support type concrete slabs may span one-way or two-way. For
supports on two sides one-way span is the only option. For supports on all sides and
about equal span in both directions, two-way slabs are better, but for unequal spans one-
way slabs are better. Deflection increases with the fourth power of span or 16 times
greater deflection for a double span. Therefore, rebars spanning the long way are
ineffective since the shorter span deflects less and carries most load. Slabs and plates
have reinforcement at the bottom of mid-span and on top of multi-span supports. The
diagrams show only one layer of reinforcement for clarity. Sections on the next page
show both top and bottom reinforcement. The following slab span/depth ratios L/d give
minimum slab depths if deflection is not checked.
One-way slabs have rebars in one direction but require some rebars to resist stress due
to temperature variation and shrinkage. The temperature reinforcing runs perpendicular
area as follows:
- 2 0 0 6
to main reinforcing and must be a minimum percentage of the concrete cross section
Grade 40
Grade 60
1 9 9 0
0.20% reinforcement
0.18% reinforcement
erle
Slab span/depth ratios L/d
i
Grade Cantilever Simple support one end cont... Both ends cont.
GS c h40
60
13
10
25
20
30
24
35
28
G
Two-way slabs need no temperature rebars. Slabs without beams require different
r i g h t reinforcement for middle strips and column strips. Column strips need more rebars since
they carry a greater load share. Two-way slabs need about 20 % less depth, but require
Copy support on all sides.

1
2
3
One-way slab on walls
One-way slab on beams
Two-way slab on walls; L/d = 36 for multiple bays
4 Two-way slab on beams; L/d = 36 for multiple bays
5 Two-way slab on columns with drop panels to resist shear at columns;
L/d = 33 for multiple bays
6 Two-way plate on columns without drop panels;
L/d = 30 for multiple bays (for moderate load; low formwork cost)
A Concrete strip of 1’ (1 m) wide assumed for slab analysis like a beam.
B Reinforcement of one-way slab
C Temperature reinforcement of one-way slab
D Reinforcement of two-way slab running both ways
E Middle strip reinforcement
F Column strip reinforcement carries a greater share of the load

Slab and plate reinforcement
Slabs and plates require bending reinforcement at zones of tensile stress, at bottom of
mid-span, top of fixed-end support, and of multi-span supports. One-way slabs require
some perpendicular reinforcement due to temperature variation and shrinkage. Thin
slabs are sometimes reinforced with welded wire fabric rather than individual bars. Shear
stress is usually resisted by concrete alone without shear reinforcement. Flat slabs
without beams have drop panels, or column caps, to resist high shear on top of columns.
Plates are similar to slabs without drop panels or column caps. They are commonly used
for relatively light loads, such as roofs.
Depth of slabs and plates, depending on span, range from
4 to 12 in (10 to 30 cm) for slabs on beams
6 to 12 in (15 to 30 cm) for flat slabs on columns; and
1
6 to 14 in (15 to 36 cm) for plates
One-way slab on beams
- 2 0 0 6
2
3
Two-way slab on beams
1 9 9 0
Two-way flat slab on columns with drop panels, column strip
erle
4 Two-way flat slab on columns with drop panels, mid-strip
5 Two-way flat slab with mushroom columns, column-strip
c h
6
i Two-way plate without drop panels, column strip
GS
A Top reinforcement at zone of negative bending
B Temperature reinforcement of one-way slab
i g h t G C
D
Bottom reinforcement in zone of positive bending
Beam
y r E Drop panel on top of column
Cop
F Mushroom panel on top of column

Rib slab
Rib slabs, also called pan joists, are one-way systems for medium spans where flat slabs
or plates would be too deep and heavy. Rib slabs reduce dead weight by eliminating
concrete between ribs, providing structural depth without bulk. The tensile steel for
positive bending is placed at the bottom of ribs and rib top and slab, resist compressive
stress like a T-beam. Given narrow spacing of ribs, the minimum slab depth is
determined not by depth/span ratio but by rebar size plus concrete cover. Long ribs may
need bracing by cross-ribs to prevent buckling. Because of one-way span, rib slabs are
not limited to square plans like two-way slabs. Rib slabs are formed placing reusable,
prefabricated pans of plastic or steel on wood boards. Ribs may be tapered near the
supports to resist the maximum support shear. Design and analysis of rib slabs is similar
to T-beams.
Rib slab dimensions:
Span/depth ratio L/d = 20-24 (d = total depth)
- 2 0 0 6
0
Maximum span = 50’ (15 m)
9
Slab depth 2.5” to 4.5” (6 to 11 cm)
1
Total depth 12” to 24” (30 to 60 cm)
9
erle
i
Rib spacing 2’ to 3’ (60 to 90 cm)
GS c hSlab on beam
Slabs on beams may be one- or two-way systems depending on proportions of spans
G
between the boundary beams. For equal spans, two-way slabs are appropriate. For
r i g h t unequal spans, one-way slabs are better and should span the shorter direction. Slabs on
beams are an intermediate solution between flat slab or plate, and rib or waffle slabs.
Copy The formwork for slab on beam is more complex and costly than for rib slabs.
The slab on beam is designed by the Direct Design Method as two-way system and as
beam-like strip as one-way system.
Slab on beam dimensions:
Span/depth ratio L/d = 30-36
Maximum span = 30’ (9 m)
Slab depth 4” to 12” (10 to 30 cm)
1 Rib slab braced by intermediary cross rib
2 Slab on beam may be one-way or two-way span

Waffle slab
Waffle slabs are two-way systems for medium spans where flat slabs or plates would be
too deep and too heavy. They reduce dead weight by two-way ribs to eliminate excess
weight between the ribs. The tensile steel for positive bending is placed at the bottom of
ribs and the rib top and slab resist compressive stress. Since negative bending reverses
the stress, waffle slabs are not efficient as cantilever with negative bending. Waffle slabs
need either solid panels on top of columns to resist shear stress or two-way beams.
Waffle slabs are formed placing re-usable prefabricated pans of plastic or steel over a
grid of wood boards.
Waffle slabs may be designed similar to the previously described Direct Design Method
like a flat slab on columns. Typical waffle dimensions: 2’ to 5’ (60 to 150 cm)
6
1 Waffle slab with a single solid panel over columns
2
3 Waffle slab supported by beams
0 - 2 0 0
Waffle slab with four solid panels over columns
1 9 9
c h i erle
h t G GS
y r i g
Cop

Special slabs
The Italian engineer Arcangeli proposed a waffle slab with curvilinear ribs that follow
isostatic lines for optimum stress distribution and a more elegant appearance. Pierre
Luigi Nervy built such a slab for a wool factory in Rome. For a tobacco factory in
Bologna, Nervi built a waffle slab with ribs wedged toward supporting beams for
increased shear capacity.
1 Slab with isostatic ribs, proposed by Arcangeli
2 Slab with isostatic ribs; for a wool factory in Rome, by Nervi
3 Waffle slab with ribs wedged to increase shear capacity
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Foundation
Foundations include building parts connected with the ground, namely slabs on grade,
basement walls, and footings supporting walls and columns. Foundations are often
exposed to groundwater and thus corrosion. Given their corrosion resistance, concrete
foundations are very common. Foundations must support various load combinations:
• Gravity live and dead load
• Lateral loads caused by earthquake, wind, and earth pressure
• Lateral thrust of arches, suspension cables, domes, shells, etc.
• Uplift caused by wind on light-weight structures, overturning of tall buildings due to
wind and seismic load; or groundwater buoyancy.
Foundation design is based on the following considerations:
• Imposed loads described above
•
•
Allowable soil pressure of various soil types
- 2
Allowable soil pressure of various soil types
0 0 6
•
•
1 9 9 0
Soil settlement (critical when soil of uneven stiffness exists on a site)
Allowable bending and shear stress of reinforced concrete foundations
erle
• Depth of frost (ground water expansion due to frost causes cracks)
• Hot climate ~ 0, moderate climate ~ 3 ft (1 m); cold climate ~ 7 ft (2 m)
c h i
Soil Capacity, approximate allowable pressures (rounded for kPa)
GS
Soil type Approximate allowable soil pressures
G
Soft clay 2 ksf 100 kPa
r i g h t Stiff clay
Sand, compacted
4 ksf
6 ksf
200 kPa
300 kPa
y
Gravel 15 ksf 700 kPa
Cop
Sedimentary rock 50 ksf 2400 kPa
Hard rock (basalt, granite) 200 ksf 9600 kPa
1 Column footings are usually square, from 1 to 2 ft (30 to 60 cm) thick with two-way
rebars at bottom and dowel bars extending into column
2 Grade beams carry uniform load (of walls) to caissons, piers, orpiles;
or distribute column loads and tie them together as shown
3 Wall footings are linear and usually about twice the width of the wall, with bottom
rebars in length direction and cross bars for wide footings
4 Mat foundations can bridge uneven settlements in variable soil and may
counteract groundwater buoyancy with appropriate mat depth
5 Pile caps distribute column load to piles and tie them together
6 Grade beams may connect pile caps for increased lateral stability

Pile and pier
Piles and piers, or caissons, are used in poor soil conditions. They have been used since
antiquity. Vitrusius wrote of wood piles driven by machinery. Piles are driven into soil but
piers are poured into excavated shafts.
Piles are driven into soft soil by a pile driver either as friction piles relying on soil friction,
or end-bearing piles to rest on firm soil or rock. Uplift loads are resisted by friction with
coefficients ranging from 0.35 to 0.6. Pile capacities are verified by resistance to the last
blows or by test loads. Piles can be vertical or inclined for lateral load. Due to capacity
limits and difficulty in precise placement, piles are in clusters of usually three or more,
spaced about 30 to 36 in (75 to 90 cm) and supporting columns at cluster centroid. Pile
caps distribute column load to piles and tie them together. Piles come in steel, wood and
concrete. Wood piles are peeled and pressure treated trunks. They need dry soil or
should be fully under water to prevent rotting. Steel piles of round pipes or H-shapes can
6
be weld-spliced to great length as they are driven. H-piles displace the least soil for easy
- 2 0 0
penetration. Concrete piles may be plain, reinforced, prestressed, precast, or site-cast.
They are the most common because of inherent corrosion resistance.
0
Pile type
Wood
1 9
Diameter
12 - 24 in
9 Max. length
80 ft
Capacity range
60 - 100 k
erle
30 - 60 cm 24 m 270 - 450 kN
i
Steel pipe 10 - 36 in 200 ft 100 -400 k
GS c hSteel H-pile
25 - 90 cm
8 - 14 in
60 m
300 ft
450 - 1800 kN
80 - 400 k
G
20 - 36 cm 91 m 360 - 1800 kN
r i g h t Concrete 8 - 36 in
20 - 90 cm
200 ft
60 m
60 - 200 k
270 - 900 kN
Copy Piers are usually shorter than piles and are used in firmer soil where pile driving is
difficult. They are cast in place against excavated soil or a steel form that is gradually
removed as the concrete is poured. Piers resting on soft soil may need a bell at the
bottom to enlarge the bearing area but those on rock have straight shafts. The bell is
formed by partially removing the form and compacting the concrete to push it outward.
Placing of piers is more precise than piles. Thus, only a single pier is needed to support
a column. Pier diameters range from 1.5 to 7 ft (0.5 - 2.1 m) with bells 2 to 3 times wider.
Capacities range from 70 to 10,000 k (300 to 45,000 kN).
1 Steel H-pile cross section
2 Wood pile cross section
3 Concrete pile cross sections
4 Piles: end-bearing pile at left; friction pile at right
5 Piers: bell pier at left; straight pier at right
6 Pile caps: common plans and cross sections

Retaining wall
Retaining walls facilitate abrupt changes in topography. They resist lateral soil pressure
and possible surcharges, such as buildings. Three types of retaining walls are mass
walls (rare today), cantilever walls of concrete and masonry. Retaining walls should have
weep holes spaced about 10 ft (3 m) with gravel backfill. They also require expansion
joints spaced about 30 ft (10 m) to prevent cracking. Retaining wall design depends on
location. At property lines the footing must point away from the property line, otherwise
footings are located to best resist overturning, using soil as ballast. Cantilever retaining
walls require footing widths of about 2/3 the height from top of footing to top of wall.
Walls with sloped backfill need footings of about 1-1/4 their height and also need a key
below the footing to help resist lateral sliding. Walls up to 6 ft (1.8 m) height require 8 in
(20 cm) width. Higher walls up to 9 ft (3 m) height require 12 in (30 cm) at their lower
6
portion. Depending on height, vertical rebars range from # 3 to # 8, spaced 8 to 32 in (20
- 2 0 0
to 81 cm) and horizontal bars are spaced about 16 in (40 cm). Retaining walls are
usually designed using equivalent fluid pressure as lateral load.
0
9 9
Mass walls resist lateral pressure by their mass or dead weight and are thus very bulky.
1
They are usually of plain concrete, but may be reinforced to reduce cracking. Mass walls
erle
are about 12 in (30 cm) wide on top and increase in width about 1/3 of the distance from
i
the top.
GS c hConcrete walls are cantilever retaining walls that resist lateral pressure by being
cantilevered from the ground. They balance overturn moments by their own weight
G
combined with soil surcharge imposed on their footing and resist sliding by lateral soil
r i g h t pressure and friction at the base. They are more expensive than concrete masonry walls
due to the expense of formwork.
Copy Concrete masonry walls are also cantilever retaining walls that resist lateral pressure
by cantilever action. They resist overturn moments by their own weight and soil imposed
on the footing and resist sliding by lateral soil pressure and friction at the base. The
footing usually requires a key below the footing to help resist lateral sliding. Concrete
masonry retaining walls are most common due to a balance of strength and economy.
1-3 Mass retaining walls
4 Concrete /CMU wall at property line with adjacent land lower
5 Concrete / CMU wall not at property line
6 Concrete / CMU wall at property line with adjacent land higher
7 Concrete / CMU wall at property line with adjacent land lower
8 Concrete / CMU wall not at property line
9 Concrete / CMU wall at property line with adjacent land higher

Prestressed concrete
The effect of prestress on concrete is to minimize cracks, reduce depth and dead weight,
or increase the span. Analogy with a non-prestressed beam clarifies this effect. In a
simply supported non-prestressed beam the bottom rebars elongate in tension and
concrete cracks due to tensile weakness. In prestressed concrete tendons (high-strength
steel strands) replace rebars. The tendons are pulled against concrete to compress it
before service load is applied. Service load increases the tension in tendons and
reduces the initial concrete compression. Avoiding tensile stress in concrete avoids
cracks that may cause corrosion in rebars due to moisture. Further, prestress tendons
can take the form of bending moments that balances the service load to minimize
deflection. This, combined with higher strength concrete of about 6000 psi (40 MPa),
allows for longer span or reduced depth in beams.
6
Pre-tensioning and post-tensioning are two methods to prestress concrete. They are
- 2 0 0
based on patents by Doehring (1886) and Jackson (1888); yet both were unsuccessful
due to insufficient stress that dissipated by creep. Doehring stressed wires before
0
9
casting the concrete, and Jackson used turnbuckles to stress iron rods after the concrete
1 9
had cured. Subsequent experiments by others led to the first successful empirical work
erle
by Wettstein in 1921 and the first theoretical study by French engineer Eugene
Freyssinet during 1920, followed by his practical development. In 1961 the US engineer
c h i
T Y Lin pioneered prestress tendons that follow the bending diagram to balance bending
GS
due to load. Lin’s method controls deflections for any desired load, usually dead load
and about half the live load. By his method, before live load is applied a beam (or slab)
i g h t G bows upward. Under full load, they deflect and under partial live load they remain flat.
Diagram 5 illustrates this for a simply supported beam.
y r 1 Simply supported beam without prestress, cracked at tensile zone
Cop
2 Prestress beam with concentric tendon deflects under service load
3 Prestress beam with eccentric tendon pushes up without service load
4 Same beam as 3 above with service load balanced at max. mid-span moment but
not elsewhere since tendon eccentricity is constant
5 Prestress beam with parabolic tendon to balance bending moment
A Bending stress: top concrete compression and bottom steel tension
B Prestress uniform due to concentric tendon
C Prestress with greater compression near tendon at bottom
D Prestress for eccentric tendon: bottom compression and top tension (tension where
tendon is outside beam’s inner third (Kern). Simply supported beam of zero end
moments has concentric tendon at ends
E Service load stress: top compression and bottom tension
F Combined stress from prestress and service load
G Combined stress with uniform distribution due to balanced moments

Tendons are high strength steel strands used in prestressed concrete. They have a
breaking strength of 270 ksi (1860 MPa) and are composed of 7 wires, six of them laid
helically around a central wire. Tendons come in sizes of 0.5 and 0.6 in (13 and 15 mm)
diameter. Both sizes are used in post-tensioned concrete but only small tendons are
used in pre-tensioned concrete which requires no bond length. The great strength of
tendons allows initial stress levels high enough to make up for loss of stress due to
creep, most notably during initial curing.
Post-tensioning begins by placing of metal or plastic tubes that house the prestress
tendons prior to the pouring of concrete. Some tendons come enclosed in the tubes, but
most are inserted after concrete has cured. Tubes prevent tendons to bond with the
concrete to allow free movement for subsequent prestress operation. Once concrete has
reached sufficient strength, the tendons are prestressed using hydraulic jacks that press
2 0 0 6
against concrete to transfer prestress into it. Short members have tension applied at one
end only but long members may require tension at both ends to overcome friction.
-
0
Several devices are available to anchor the ends of post-tensioned tendons into
9 9
concrete. One such device is a conical wedge that holds tendons by mechanical friction
1
between the tendon and a rough surface of the device. Post-tensioning is usually done
erle
at the building site. It allows tendons to take any form desired to balance bending
i
moments induced by service load.
GS c h1
2
Tendon cross section.
Conical tendon lock rests against steel sleeve (not shown) inserted into concrete
G
and squeezes the tendon and hold it by friction
rig h t 3
4
Beam with straight eccentric tendon below neutral axis
Parabolic tendon emulate bending moment to reduce deflection
Co p y 5
6
7
Multiple parabolic tendons with offset anchors
Continuous beam with tendons emulating bending moment distribution
Continuous slab with tendons emulating bending moment distribution

Pre-tensioning stresses tendons between abutments in a precast concrete plant. Once
concrete has sufficient strength (with the help of steam in about 24 hours) tendons are
cut off at the abutments to transfer prestress into the concrete. After cut off, tendons are
anchored by friction with the concrete at both ends. Since abutments are difficult to
secure at construction sites, pre-tensioning is usually done in a precast plant. Similar
pieces may be laid in a row and cast together requiring only two abutments per row. Pre-
tensioning is simplest with straight tendons, but some approximate curves that emulate
bending moments are possible. They require temporary tie-downs to be cut off along
with tendons after curing is complete.
Pre-tensioned members must be carefully handled during transportation to avoid damage
or breakage. Since the reinforcement is designed for a given load direction, any
reversed load may result in overstress and possible breakage. To avoid this they must
6
be placed on the truck in the same position as in their final installation. Also, to avoid
1 Beam with tendons anchored to abutments
0 - 0 0
breakage of corners, it is advisable to provide corners with chamfers.
2
2
9 9
Beam with tendons cut off to transfer prestress into concrete after it has reached
1
sufficient strength
erle
3 Beam with tendon tie-down to approximate bending moment curves
i
4 Row of pre-tensioned members with tendon anchors at ends only
GS c h5
For members like walls and columns with possible bending in any direction,
tendons may be placed at the center
Row of pre-tensioned beams with tendon tie-down to approximate bending
h t G moments of service loads
rig
A Close approximation of parabolic bending moment by tendon shape
C o p y B Temporary tendon tie-downs to approximate bending moments are cut off after
tendons are cut from abutments

Precast concrete
Precast concrete comes in a wide variety of shapes for both structural and architectural
applications. Presented are structural systems and members: floor and roof members,
columns, and walls. Though precast members may be of ordinary concrete, structural
precast concrete is usually prestressed. The primary reinforcement of prestressed
concrete is with tendons, yet normal rebars are often used as stirrups to resist shear.
Rebars are also added for different loads during transportation and erection. Compared
to site-cast concrete, precast concrete provides better quality control, repeated use of
formwork, faster curing with steam, and concurrent operations while other site work
proceeds. The advantages must offset the cost of transportation to a construction site.
Precast concrete is similar to steel framing by allowing preparatory site work to be
concurrent, yet it has the advantage to provide inherent fire resistance. Steel on the
2 0 0 6
other hand, has lower dead weight, an advantage for seismic load that is proportional to
dead weight. To reduce high costs of formwork the number of different precast members
-
0
should also be reduced; yet this objective must be balanced by other considerations. For
9 9
example, fewer parts may result in a monotone and uninspired design. Combining
1
precast with site-cast concrete may satisfy economy as well as aesthetic objectives.
erle
Precast framing allows many variations, both with and without site-cast concrete. A few
c h i
typical examples are presented. They are possible with columns of several stories,
GS
limited primarily by transportation restrictions. The capacity of available cranes could
also impose limitations. In such cases, columns should be spliced near mid-height
h t G between floors where bending moments from both gravity and lateral loads are zero.
rig
1 T-columns with deep spandrel beams support floor and roof slabs. Shear
y
connections between adjacent beams combine them to moment frames
C o p 2
to resist lateral as well as gravity loads
Frames of split columns and deep spandrel beams support floor and roof slabs for
gravity and lateral loads. Shear connections at adjacent split columns tie the
frames together for unified action
3 T-columns with normal spandrel beams support floor and roof rib slabs
Shear connections between adjacent beams combine them to moment frames to
resist lateral as well as gravity loads
4 Tree-columns with beam supports allow flexible expansion. Twin beams allow
passage of services between them. Lateral load resistance must be provided by
shear walls or other bracing
5 Rib slab or double T’s supported on site-cast frame
6 U-channels with intermittent skylights supported on site-cast frame

Floor and roof members span primarily horizontally to carry load in bending to
supporting columns or walls. Some are designed for bridge structures but also used in
buildings. A primary objective is to keep the dead weight low yet providing relatively long
spans. To this end long-span members have ribs or hollow cores to reduce weight yet
seeking optimal synergy for the tensile and compressive properties of steel and concrete,
respectively. Precast members are usually covered with site-cast concrete of about two
inch (5 cm) to provide a smooth surface and bond individual panels together. The
following dimensions are approximate and vary by precast plant.
Item B D L L/D
1 Solid plank 2-4 ft 4-8 in 10-30 ft 25-40
0.6-1.2 m 10-20 cm 3-10 m
2 Hollow plank 2-8 ft 6-12 in 20-50 ft 30-40
6
0.6-2.4 m 15-30 cm 7-15 m
3 Single T
0
4-8 ft
1.2-2.4 m
12-48 in
30-120 cm
-2 0 0 20-120 ft
7-36 m
20-40
9
4 Double T 4-8 ft 12-48 in 20-120 ft 20-40
5 I-beam
r l e 1 9 1.2-2.4 m 30-120 cm
20-48 in
6-36 m
20-120 ft 15-30
i e
50-120 cm 6-36 m
G Sc h6
7
L-beam
Inverted T
20-48 in
50-120 cm
20-48 in
20-60 ft
6-24 m
20-60 in
10-20
10-20
i g h t G 50-120 cm 6-24 m
p y r
Co

Precast columns must support beam gravity load and prevent sliding under lateral load. Concealed moment resistant joints
This is possible by steel dowels, grouting, bolting, welding, or tie-rods. Dowel bars are 1 Concealed beam/column joint
joint by grouting after erection. Bolting and welding connect metal plates welded to A Steel angle cast into beam end
rebars of joining members. Bearing pads prevent overstress due to uneven surface of B Wide-flange steel haunch cast into column
beam or column. Precast columns several stories high, limited by transportation C Rebars welded to steel haunch
constraints are spliced at the mid-height of a story where moments are zero. 2 Moment resistant beam/beam joint (or similar beam/column joints)
1 Column with semi-corbel recessed into beam for flush bottom face A Dry-pack grout
2 Column with two beams connected by steel dowels B Post tensioned rod extending through column
3 Column with one-sided corbel to support beam C Anchor plate
4 Column with two-sided corbels to support beams D Pocket for tensioning jack
5 Column splice with recessed metal brackets E Grouted pocket after completion of post tensioning
6 Column to footing connection with steel plate and anchor bolts F Bearing pad
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Precast wall
Walls can be precast in many sizes and configurations: small modular panels; walls
extending entire spaces; complete walled boxes (hotel rooms); single story walls; and
multi-story walls. The choice of size and configuration depends on means of construction
and limitations imposed by transportation requirements. Exhaustive coverage is beyond
the scope of this book. Only some basic conditions are presented. Precast walls may be
prestressed or with normal reinforcement that would be similar to site-cast concrete walls
presented before. For thick walls two layers of tendons are placed near the wall faces.
For most walls single layers are placed at the center of walls to resist buckling and
bending in any direction. For walls that continue over several stories concrete floors and
roofs are supported by concrete corbels, steel and wood by steel and wood ledgers,
respectively. Walls extending from floor to ceiling are anchored to floors or roofs and
adjacent walls by welded or bolted steel brackets, or by dowel bars inserted in grout or
6
site-cast concrete. Exterior walls may be temporarily stabilized by dowel bars that are
that end and is filled after the wall is done.
0 - 2 0 0
tied or welded to dowel bars extending from floor slabs. A slot in the slab is left open to
1
2
9 9
Roof support by exterior wall corbel with welded brackets
1
Roof support by interior wall with bolted and welded brackets
erle
3 Floor support by exterior wall corbel with welded brackets
i
4 Floor support by interior beam with bolted bracket
GS c h5
6
Exterior footing with wall tied to floor by overlapping dowels. A slot in floor slab is
filled after dowels are tied or welded together
Interior column footing
i g h t G A Metal reglet for flashing, inserted in concrete wall
r
B Roofing membrane on rigid thermal insulation or site-cast concrete
Copy C
D
E
F
Metal plate welded to connect brackets in wall and floor or roof
Precast roof slab (hollow core or other)
Bearing pads of neoprene or rubber to distribute load evenly
Grouted dowel bars to connect adjacent floor slabs
G Site-cast concrete top for smooth floor and to join precast slab panels
H Steel bracket to tie walls of adjacent floors together
I Pocket for bolting bracket (grouted after completion)
J Corbel for floor and roof (or ledger for wood or steel floors and roofs)
K Concrete beam, precast or site-cast
L Slot in floor slab to tie or weld wall to floor dowels (filled subsequently)
M Wall and floor dowels tied or welded together to anchor wall to floor
N Gravel to support floor slab in connecting slot

Tilt-up concrete
Tilt-up concrete is a precast technique performed on the building site. Due to relatively
low cost, tilt-up construction is popular for industrial buildings and warehouses, but is
adaptable to other building types. Wall panels are cast on top of the floor slab under
construction to eliminate most formwork. The floor slab must be smooth and flat and
carefully treated with a bond-breaking compound to prevent wall panels to bond with the
slab. Depending on the wall finish desired, panels may be cast with the outside face
facing up or down. The latter is more common and is the only logical option when
corbels or other projections for floors or roofs are needed on the inside. Electrical wiring
and similar items needed, including lifting and temporarily bracing inserts, are installed
2 0 0 6
prior to concrete. Panels may be up to 30 ft (9 m) wide and 60 ft (18 m) high, but they
are usually much smaller. The lifting capacity of available cranes must be considered in
-
0
selecting panel sizes. The crane capacity should be about twice the panel weight to
9 9
account for initial inertia and panel bonding with the floor slab. Once the panel has
1
reached sufficient strength, after about seven days, the lifting process begins. Panels
erle
must be designed not only for service load but for any possible load during lifting and
i
erection. After erection, the wall panels are braced with telescopic steel braces that allow
GS c hproper alignment. The panels are connected to the floor slab by overlapping dowel bars
that extend from wall panels and floor slab. Dowels are tied or welded together to anchor
panels to the floor for initial stability until roof or floor diaphragms are in place. Those
i g h t G diaphragms transfer lateral loads to wall panels that act as shear walls parallel the load.
r
A Wall panel installed with temporary bracing
Copy B
C
D
E
Wall panel being tilt-up and lifted into the footing
Wall panel on the floor slab ready for tilt-up
Steel braces with telescopic spindle for panel alignment
Floor slab dowel bars to be tied or welded to wall dowels
F Wall panel dowel bars to be tied or welded to floor dowels
G Possible dowel bars at panel edges to tie panels together
H Steel ledger supports steel roof (wood ledgers or concrete corbels
are used for wood or concrete roofs)

Tilt-up details are similar to other precast wall details but some are unique to tilt-up
construction. Wall panels may be connected by various butt joints or cast-in-place
concrete splices or pilasters or by precast columns. Only some common details are
presented. Walls may have concrete corbels, steel or wood ledgers, for concrete, steel
and wood floors or roofs, respectively. Reinforcement, except for connecting dowels, is
not shown for clarity.
1 Wall corner with cast-in-place concrete splice
2 Wall corner with mitered butt joint and sealant
3 Wall joint with cast-in-place concrete splice
4 Wall joint with cast-in-place concrete pilaster
5 Wall joint with grouted key in precast double pilaster
6 Wall joint with grouted key in butt joint
6
7 Roof supported by exterior wall corbel with welded brackets
8
9
- 2 0
Roof supported by exterior wall with steel ledger
Roof supported by exterior wall with wood ledger
0 0
9
10 Exterior footing with wall tied to floor by overlapping dowels. A floor slab slot is
9
filled after dowels are tied or welded together
1
erle
11 Exterior footing after floor slab is completed
i
A Tilt-up wall panel
GS c hB
C
D
Cast-in-place concrete splice
Dowel bars extending from wall panels into concrete splice
Grouted concrete shear key ties panels together
i g h t G E
F
Silicone sealant with backer rod
Metal reglet for flashing, inserted in concrete wall
y r G Metal plate welded to connect wall and roof brackets
Cop
H Roofing membrane on rigid insulation or concrete topping
I Bearing pads of neoprene or rubber to distribute load evenly
J Corbel to support roof or floor
K Steel angle ledger to support metal roof or floor
L Wood ledger to support wood roof or floor
M Anchor bolt to connect steel or wood ledger to wall panel
N Slot in floor slab to tie or weld wall to floor dowels
O Wall and floor dowels tied or welded together to anchor wall to floor
P Gravel to support floor slab in connecting slot

Projects
Hellas research foundation, Crete
Architect: Panos Koulermos
Engineer:Technical offices of the Hellas Research Foundation and the Greek Ministry of
Industry, Research and Technology
This research foundation has four building blocks, clockwise from top left: administration
and mathematics; physics and laser research; molecular biology; and a plaza over
research offices. A vaulted gallery, reminiscent of the vaults in nearby Heracleon, is
supported by columns that flank a gallery to link the various elements, evoking memories
of classical Greek architecture. Moment resisting reinforced concrete frames supports
two-way concrete slabs between its beams. The frames are based on a module that
varies from 4.8 to 6m (16 to 20 ft) in response to program needs and is combined with
2 0 0 6
shear walls for increased resistance to lateral load. Concrete masonry partitions provide
further resistance. Koulermos explored the plastic qualities of concrete to articulate
-
0
facades in response to program needs. The oblique research wings integrate in a 1.2 m
9 9
(4 ft) deep envelope zone exposed columns and u-shaped walls for sun control and as
1
mechanical chases that facilitate changing needs. The plaza block features freestanding
erle
columns to frame dramatic views and support sun-shading devices. Articulation of the
i
administration wing facade reflects interior functions. Facade elements of research
GS c hwings, next to typical floor framing, are shown below.
i g h t G
y r
Cop

Terrace Homes Taipei, China (1980)
Architect: G.G. Schierle
Engineer: Kuan Wai Yen
This 200 unit community on a hillside with dramatic valley view includes eight types of
terrace homes for various needs and local topographies. Most units are in groups of
four; two of them with access from below with street level parking, living above and
bedrooms on the third floor. The other two units, with access from above, are in
reversed order. Both types have two large terraces with party walls and planters for
privacy.
Moment resistant concrete frames, based on a grid in the range of 3 to 4 m (10 to 13 ft)
in response to space needs, support two-way concrete slabs and are vertically aligned
for straight load paths. Masonry walls add stiffness to minimize movement in wind or
6
moderate earthquakes, whereas frames provide ductility for fail-safe performance should
0 - 0 0
brick walls fail in severe earthquakes, a proven effective combination in seismic regions.
2
Two-way reinforced concrete slab on beams, all site-cast
B
C
9
Concrete slab on grade
1
Retaining wall with waterproofing
9
erle
D Concrete beam
c h i
h t G GS
y r i g
Cop

Lloyd’s of London (1986)
Engineer: Ove Arup
Loyd’s of London, a society of syndicated underwriters, needs a market place known as
“The room”. The program required a facility for the twenty-first century, including an
underwriter room three times larger than the old one. Given the location in London’s
financial district, an area of mostly small and winding streets, the design is in response to
this urban context as well as the program. A permanent structure for the central space,
expected to last, is flanked by service towers, expected to adapt with changing
technology. The central space is an efficient box with an atrium covered by vaulted truss.
Roof terraces step from 12 stories facing a high-rise on the North to six stories on the
South. The service towers with stairs and elevators respond to small scale urban
context. Located on the outside they allow continued operations during future changes.
- 2 0 6
Reinforced concrete columns, spaced 35x59 ft (11x18 m), support the main structure.
0
One-way ribs at 1.8 m (6 ft) and cross-ribs act as diaphragm to carry lateral load to
1 9 9 0
vertical bracing. Cross ribs distribute rib load to adjacent ribs
Access floors on stubs provide space for service distribution
erle
C Precast, prestressed concrete channel beams
D Cantilever brackets support beams
c h
E
i Diagonal bracing on six façade bays resists lateral loads
h t G GS
y r i g
Cop

Cafeteria, state college Hayward, California (1967)
Architect: Worley K Wong and Associates
Engineer: Eric Elsesser
This cafeteria was built with two structural precast concrete components: Channel beams
supported by channel columns; though the beams vary in lengths and width. A large
interior atrium links the two levels and provides visual relieve to the strictly modular
structure. The columns cantilever from underground foundations to resist lateral load in
bending as well as gravity axial load. Columns step back at floor and roof levels to
provide supporting shoulders for beams. Thus, the columns are largest at the base
where moments from lateral loads are greatest. Roof beams are slightly wider than floor
beams due to column setbacks. The void of channel columns, and the aligned space
between beams, house electrical and mechanical services and recessed folding
partitions. On-site concrete provides the finish over and bridges the gap between beams.
- 2 0 0 6
Beams are connected to columns by inserts, welded together after erection. Neoprene
pads provide smooth load transfer from beams to columns. Most of the precast concrete
0
has exposed aggregate finish; only the inside of channel beams is smooth and painted
9 9
white for a striking contrast and to enhance natural lighting by reflection.
1
c h i erle
h t G GS
y r i g
Cop

Max-Eyth school Schöntal, Germany Hampshire national building, Culver City, California
Architect: P. M. Kaufmann Architect: James Tyler
Engineer: W. Böck Engineer: Dimitry Vergun
Located on the edge of the Jagst valley with its rich history, the school is named after the This two-story facility of three 60x100 feet units is designed on a 25/25 feet module. Tilt-
poet-engineer Max-Eyth. The terracing follows the natural grade and adjacent vineyards. up concrete panels on the long sides are joined to steel moment frames on the short
Clear story light floods the central atrium and stair. The atrium provides visual continuity. sides to resist gravity load and lateral load in length and width direction, respectively.
Terracing creates dynamic space composition. The structural grid provides vertical Interior steel columns carry gravity load only. The 6 inch tilt-up wall panels, 25/25 feet
continuity for load paths and installations. A concrete moment frame of 8.4/8.4/3.6 m are spliced with poured-in-place concrete, and welded to steel columns at four corners.
resists gravity and lateral loads. Exposed two-way joists, spaced 2.4 m, span the square Steel truss joists support floor and roof metal decks. The tilt-up panels were poured on
modules and support a two-way concrete slab. The exposed concrete frame and precast the concrete floor, adjacent to their erected positions. The panel’s outside was
exterior wall panels are contrasted by wood partitions sandblasted for a textured gravel finish.
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop

Material
24 Tent membranes have been around since ancient history, notably in nomadic societies.
However, contemporary membrane structures have only evolved in the last forty years.
Structural membranes may be of fabric or cable nets. Initial contemporary membrane
structures consisted of
• Natural canvass for small spans
• Cable nets for large spans
Industrial fabric of sufficient strength and durability was not available prior to 1970.
Ca ble a nd Fa bric Contemporary membrane structures usually consist of synthetic fabric with edge cables
or other boundaries. Cables and fabric are briefly described.
Fabric for contemporary structures consists of synthetic fibers that are woven into bands
6
and then coated or laminated with a protective film
Common fabrics include:
• Polyester fabric with PVC coating
0 - 2 0 0
•
•
9
Glass fiber fabric with PTFE coating
1 9
erle
Glass fiber fabric with silicon coating
• Fine mesh fabric, laminated with PTFE film
c h i
Fabric properties are tabulated on the next page. Foils included are only for very short
GS
spans due to low tensile strength. Unfortunately the elastic modulus of fabric is no longer
provided by fabric manufacturers, though it is required for design and manufacture of
h t G fabric structures. The elastic modulus of fabric is in the range of:
rig
E = 2000 lb/in, 11492 kPa/m to
C o p y E = 6000 lb/in, 34475 kPa/m

Cables may be single strands or multiple strand wire ropes as shown on following pages.
Cables consist of steel wires, protected by one of the following corrosion resistance:
• Zinc coating (most common)
• Hot-dip galvanizing
• Stainless steel (expensive)
• Plastic coating (used at our cable nets at Expo64 Lausanne)
Depending on corrosion protection needs, zinc coating comes in four grades: type A,
type B (double type A), type C (triple type A), type D (four times type A). Cables are
usually prestressed during manufacture to increase their stiffness.
Elastic modulus of cables:
E = 20,000 ksi, 137900 MPA (wire rope)
E = 23,000 ksi, 158,585 MPa (strand > 2.5 inch diameter)
E = 24,000 ksi, 165,480 MPa (strand < 2.5 inch diameter)
24-1 MATERIAL Cable/Fabric

Fabric
Type Makeup Common use Tensile strength Fire rating UV light resistance Translucency Durability
++ incombustible ++ very good
+ low flammability + good
0 none
Coated fabric* Polyester fabric Permanent + mobile 40 to 200 kN/m + + 0 to 25 % 15 to 20 years
PVC coating Internal + external 228 to 1142 lb/in
Coated fabric* Glass fiber fabric Permanent 20 to 160 kn/m ++ ++ 4 to 22 % > 25 years
PTFE coating Internal + external 114 to 914 lb/in
Coated fabric Glass fiber fabric Permanent 20 to 100 kN/m ++ ++ 10 to 20 % > 20 years
6
Silicone coating Internal + external 114 to 571 lb/in
Laminated fabric* Fine mesh fabric Permanent 50 to 100 kN/m ++ ++
0 - 2 0 0
35 to 55 % > 25 years
9
Laminated with Internal + external 286 to 571 lb/in
Foil
PTFE film
PVC foil Permanent internal 6 to 40 kN/m 0
rle 1 9 + Up to 90 % 15 to 20 years
i e
Temporary external 34 to 228 lb/in internally
Foil* Flouropolymer foil

ETFE
Permanent
Internal + external
G S c h
6 to 12 kN/m
34 to 69 lb/in
++ ++ Up to 96 % > 25 years
h t G
rig
Coated or PTFE fabric Permanent + mobile 40 to 100 kN/m ++ ++ 15 to 40 % > 25 years
y
uncoated fabric* (good qualities Internal + external 228 to 571 lb/in
C
Coated or
o p
uncoated fabric*
for sustainability)
Flouropolymer
fabric
Permanent + mobile
Internal + external
8 to 20 kN/m
46 to 114 lb/in
++ ++ Up to 90 % > 25 years
* Self-cleaning properties Maximum fabric span*

Tensile strength Maximum span
SI-to-US unit conversion: 500 lb/in 60 ft
1 kN/m = 5.71 lb/in 1000 lb/in 120 ft
* Assuming:
Live load = 20 psf, 956 Pa (wind or snow)
Safety factor = 4
Fabric span/sag ratio = 10

Cables
Cables may be of two basic types and many variations thereof. The two basic types are
strands and wire ropes.
Strands have a minimum of six wires twisted helically around a central wire. Strands
have greater stiffness, but wire ropes are more flexible. To limit deformation, strands are
usually used for cable stayed and suspension structures.
Wire ropes consist of six strands twisted helically around a central strand. They are used
where flexibility is desired, such as for elevator cables.
Metallic area, the net area without air space between wires, defines the cable strength
and stiffness. Relative to the gross cross section area, the metallic area is about: 70% for
strands and 60% for wire ropes. To provide extra flexibility, some wire ropes have
central cores of plastic or other fibers which further reduce the metallic area.
1 Strand (good stiffness, low flexibility)
2
E = 22,000 to 24,000 ksi; 70% metallic
Wire rope (good flexibility, low stiffness)
- 2 0 0 6
0
E = 12,000 to 20,000 ksi; 60% metallic
Cable fittings
Cable fitting for strands and wire ropes may be of two basic types: adjustable and fixed.
1 9 9
erle
Adjustable fittings allow to adjust the length or to introduce prestress by shortening. The
amount of adjustment varies from a few inches to about four feet
3 Bridge Socket (adjustable)
c h i
GS
4 Open Socket (non-adjustable)
G
5 Wedged Socket (adjustable)
6
A
Anchor Stud (adjustable)
Support elements
r i g h t
y
B Socket / stud
Cop
C Strand or wire rope

1 Closed socket for strand or wire rope
Length about 10 times cable size
Width about 3 to 4 times cable size
2 Open socket for strand or wire rope

Length about 10 times cable size
Width about 3 to 4 times cable size
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop 3 Threaded stud fir strand
swaged fitting
Adjustment about 2 to 3 time cable size
4 Composite socket with threaded stud

Customized adjustment
5 Threaded socket for strand or wire rope

Limited adjustment for ½” to 4” cable size

1 Cable-to-cable connection with integral strand fitting
2 Cable-to-cable connection with wire rope thimble
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS
y r i g
Cop
3 Open socked connection, perpendicular
Trapezoidal gusset plate for synergy of form and reduced weld stress
4 Open socked connection, angled

Sloping gusset plate for synergy of form and uniform weld stress distribution

Mast / cable details
The mast detail demonstrates typical use of cable or strand sockets. A steel gusset plate
usually provides the anchor for sockets. Equal angles A and B cause equal forces in
strand and guy, respectively.
A Mast / strand angle

B Mast / guy angle
C Strand
D Guy
6
E Sockets
F
G
Gusset plates
Bridge socket (to adjust prestress)
0 - 2 0 0
9
H Foundation gusset (at strand and mast)
I Mast
1 9
c h i erle
h t G GS
y r i g
Cop

Production process
Fabric pattern
To assume surface curvature, fabric must be cut into patterns which usually involve the
following steps:
• Develop a computer model of strips representing the fabric width plus seems
• Transform the computer model strips into a triangular grids
• Develop 3-D triangular grids into flat two-dimensional patterns
The steps are visualized ad follows:
1 Computer model with fabric strips

2 Computer model with triangular grid
6
2 Fabric pattern developed from triangular grid
Pattern cutting
0 - 2 0 0
9
Cutting of patterns can be done manually of automatic.
1 9
The manual method requires drawing the computer plot on the fabric
erle
The automatic method directs a cutting laser or knife from the computer plot
c h i
Note:
GS
For radial patterns as shown at left, cutting two patterns from one strip, juxtaposing the
wide and narrow ends, minimizes fabric waste.
ig h t G Pattern joining
p y r Fabric patterns are assembled by one of three methods:

• Welding (most common)
Co
1
• Sewing
• Gluing
Edge cables
Unless other boundaries are used, edge cables are added, either embedded in fabric
sleeves or attached by means of lacing.
Fabric panels
For very large structures the fabric may consist of panels that are assembled in the field,
usually by lacing. Laced joints are covered with fabric strips for waterproofing.
2 3

Cable/ fabric details (1 to 6 Frei Otto details)
1 Continued cable over mast

U-bolt connect cable to mast
2 mast top with continuing membrane edge cable

U-bolts connect cables to mast
3 Mast cross section

Three pipes joined by plate bars
4 Edge cable/ fabric corner

Twin triangular plates join edge cables at fabric corner
5 Cable clamp cross section
- 2 0 0 6
6 Fabric corner
9
Cable transfer at fabric corner
1 9 0
erle
7 Edge cable/ membrane sleeve
c h i
GS
8 Membrane laced to edge cable
i g h t G
y r
Cop

1, 2 Pneumatic cushions joint to space truss of Osaka Festival Plaza
- 2 0 0 6
1 9 9 0
c h i erle
h t G GS 3 Pneumatic cushion connected to pipe with synthetic gasket
y r i g
Cop
4 Pneumatic cushion attached to concrete with twin plates and synthetic gaskets
5 Membrane attached to pipe
6 Membrane attached to concrete with twin plates

Projects
US Pavilion, Expo 70, Osaka
Architect: Davis, Brody, Chermayeff, Geismar, De Harak
Engineer: David Geiger
The pneumatic structure was supported by diagonal cables
- 2 0 0 6
1 9 9 0
c h i erle
GS
1 Fabric / cable detail
ig h t G 2 Cable crossing clamp
p y r
Co

Watts Towers Canopy
Architect: G G Schierle with Joe Addo
Engineer: ASI
A transparent membrane suspended from radial cable trusses is designed to provide sun
shading for occasional performances at the Watts towers. The crescent-shaped roof
follows the crescent-shaped seating below. The cable trusses minimize bulk for optimal
view of the towers and facilitate fast erection and removal at annual events. The truss
depth provides desired curvature for the anticlastic membrane panels. Two membranes
provide shading for spectators and performers over the respective areas. The
architectural design is shown below. The final computer drawings are shown at right.
1 Strut top
6
2 Fabric corner
A
B
Top chord strand
Diagonal strand
0 - 2 0 0
9
C Fabric attachment
D
9
Metal plate at fabric corner, adjustable to induce prestress
1
erle
E Edge cable
F Edge webbing
c h i
h t G GS
y r i g
Cop

- 2 0 0 6
1 9 9 0
ie r l e ,
S c h
i g h t GG
p y r
Co
E-2 APPENDIX E Design Tables

- 2 0 0 6
1 9 9 0
ie r l e ,
S c h
i g h t GG
p y r
Co
E-3 APPENDIX E Design Tables

About the book About the author
Structures not only support gravity and other loads, but are essential to Professor Schierle, FAIA, has PhD and Master of Architecture degrees
define form and space. To design structures in synergy with form and from UC Berkeley and a Dipl-Ing degree from Stuttgart, Germany. He was
space requires creativity and an informed intuition of structural principles. founding Director of USC’s Graduate Program of Building Science and
The objective of this book is to introduce the principles as foundation of teaches structures at the USC School of Architecture. Prior to USC he
creative design and demonstrate successful application on many case taught at UC Berkeley and the Stanford University. He has been Visiting
studies from around the world. Richly illustrated, the book clarifies Professor at UCLA and EPFL Lausanne, lectured at AIA National
complex concepts without calculus yet also provides a more profound Conventions and these universities: Arizona, Carnegie-Mellon, Harvard,
understanding for readers with an advanced background in mathematics. MIT, Utah, Braunschweig, Delft, EPFL, Stuttgart; Mexico; and Sydney. He
The book also includes structural details in wood, steel, masonry, has received several grants from the National Science Foundation, the
concrete, and fabric to facilitate design of structures that are effective and Department of Housing and Urban Development, and FEMA for research
elegant. Many graphs streamline complex tasks like column buckling or on seismic safety. His research on lightweight structures and seismic
design for wind and seismic forces. The graphs also visualize critical safety is widely published. Dr. Schierle chaired the architectural license
issues and correlate US with metric SI units of measurement. These examination on structures and serves on the Fabric Architecture advisory
features make the book useful as reference book for professional board. He also served on the Journal for Architectural Education Editorial
architects and civil engineers as well as a text book for architectural and Board and the Fabric Structures Awards Jury. His architecture practice
civil engineering education. The book has 613 pages in 24 chapters. includes major projects in America, Asia, and Europe.
http://www.usc.edu/structures http://www-rcf.usc.edu/~schierle
Structures in Architecture G G Schierle

Architectual Structures 1 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Architectual Structures 1 PDF

Uploaded by

Copyright:

Available Formats

Structures in Architecture G G Schierle

Copyright © G G Schierle 1990-2006. All rights reserved

This book includes material from the following sources:

International Building Code 2003:

PART IV: HORIZONTAL SYSTEMS

19 Suspended high-rise Appendix A: Beam Formulas

rig h t Appendix E: Design Tables

22-18 Concrete masonry

2-1 BACKGROUND Load

h t G full beam but LL only on part of it

C o p y • Seismic load (earthquake load)

Other load issues introduced::

2-2 BACKGROUND Load

c h i7.77 Shingles 3 144

h t G 556 8.91 Industrial fabric (PVC , fiber glass) 1-2 48-96

2-3 BACKGROUND Load

Schools Classrooms 40 1.92 12 < Lr < 20

2-4 BACKGROUND Load

GS c hshear as percentage of mass is approximately:

y r • ~ 20 to 30 % for stiff shear walls

2-5 BACKGROUND Load

r i g h t • Method 2: Projected area method

2-6 BACKGROUND Load

The following examples illustrate tributary load and load path

r i g h t 2 Two beams / three columns

Copy Uniform beam load w = 2 klf

2-7 BACKGROUND Load

i g h t G Floor Σ 200 psf

Copy Column reactions at level 2

2-8 BACKGROUND Load

Load path: plywood deck / joist / beam / columns

2-9 BACKGROUND Load

rig h t w = 80 psf x 10’

Co p y w = 150 pcf x 2’x 1’

f = P/A = 2500 # / (1’x2’) f =1250 psf

2-10 BACKGROUND Load

2-11 BACKGROUND Load

Find load path and tributary load

rig h t Level 1 diaphragm

2-12 BACKGROUND Load

3 • Structural design for:

• Rupture length (material properties, i.e., structural efficiency)

3-1 BACKGROUND Basic Concepts

defined by the Moment of Inertia, I = b d3 / 12

r i g h t 10 boards, glued, S =1000/5 S = 200

3-2 BACKGROUND Basic Concepts

Rapture length defines material efficiency as strength / weight ratio:

3-3 BACKGROUND Basic Concepts

3-4 BACKGROUND Basic Concepts

3-5 BACKGROUND Basic Concepts

3-6 BACKGROUND Basic Concepts

3-7 BACKGROUND Basic Concepts

3-8 BACKGROUND Basic Concepts

3-9 BACKGROUND Basic Concepts

c h i Architectural criteria apartments/hotels

Copy Structural criteria

Structural criteria Ductile, absorbs seismic Tall structures need

Structural criteria Very stiff, good for Stiffness increases seismic

3-10 BACKGROUND Basic Concepts

3-11 BACKGROUND Basic Concepts

3-12 BACKGROUND Basic Concepts

p y rig Deformations reverse under reversed load

3-13 BACKGROUND Basic Concepts